Skip to main content
main-content

Über dieses Buch

Due to the complexity, and heterogeneity of the smart grid and the high volume of information to be processed, artificial intelligence techniques and computational intelligence appear to be some of the enabling technologies for its future development and success. The theme of the book is “Making pathway for the grid of future” with the emphasis on trends in Smart Grid, renewable interconnection issues, planning-operation-control and reliability of grid, real time monitoring and protection, market, distributed generation and power distribution issues, power electronics applications, computer-IT and signal processing applications, power apparatus, power engineering education and industry-institute collaboration. The primary objective of the book is to review the current state of the art of the most relevant artificial intelligence techniques applied to the different issues that arise in the smart grid development.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Fractional-Order PID Controller Optimized by SCA for Solar System

In this paper, the maximum power point tracking (MPPT) technique is enhanced by implementing fractional-order proportional-integral derivative (FOPID) to improve the output of DC-DC boost converter of solar system and justified by competing alongside perturb and observe (P&O) and PID-based MPPT technique. The gain variables of FOPID and PID controller highly influenced the performance of the system. Sine Cosine Algorithm (SCA) is a novel technique to explore the best couple of gain factors of both PID and FOPID controllers to provide appropriate gate pulse to converter to improve performance of the output response. In this paper, the performance of the photo-voltaic (PV) system is enhanced by conceding the oscillation, time response, settling time and maximum values of voltage, current and power of the system by adopting FOPID controller.

Raj Kumar Sahu, Binod Shaw, Jyoti Ranjan Nayak

Chapter 2. LVRT Capability Improvement in a Grid-Connected DFIG Wind Turbine System Using Neural Network-Based Dynamic Voltage Restorer

The wind power plant is one of the fastest growing electrical power sources. The integration of wind turbine into the power grid originates various power quality problems. Low-voltage ride through (LVRT) capability improvement is impregnating exigency in recent power quality issue, which is acquainted with various renewable power generation resources like solar PV and wind power plant. Dynamic voltage restorer (DVR) is a custom power device (CPD), which is connected in series with the electrical system, and it is a contemporary and efficient CPD expended in the distribution system to curb the power quality problems. In this paper, LVRT capability of doubly fed induction generator (DFIG) wind turbine system connected to a power grid is enhanced by means of DVR. LVRT capability is improved by the introduction of neural network controller. This controller fulfils the various grid code requirements. Thus, the performance of the DVR becomes far superior by the robust control technique. The simulation results were compared with conventional PI controller and the proposed artificial neural network (ANN) controller. It is proved that the proposed system raises the reliability of grid-connected DFIG wind turbine system.

Arun Kumar Puliyadi Kubendran, L. Ashok Kumar

Chapter 3. Detection and Classification of Power Quality Events Using Wavelet Energy Difference and Support Vector Machine

Detection and classification of power quality events (PQE) to improve the quality of electric power is an important issue in utilities and industrial factories. In this paper, an approach to classify PQE with noise based on wavelet energy difference and support vector machine (SVM) is presented. Here PQE signals are decomposed into ten layers by db4 wavelet with multi-resolution. Energy differences (ED) of every level between PQE signal and standard signal are extracted as eigenvectors. Principal component analysis (PCA) is adopted to reduce the dimensions of eigenvectors and find out the main structure of the matrix, which forms new feature vectors. Then these new feature vectors are divided into two groups, namely training set and testing set. The method of cross-validation is used for the training set to select the optimal parameters adaptively and construct the training model. Also, the testing set is substituted into the training model for testing. Finally, the proposed method results are compared with S-transform (ST)- and Hilbert-Huang transform (HHT)-based PQE classification to verify the accuracy of classification. The results demonstrated show that the proposed method has high accuracy, strong resistance to noise, and fast classification speed and is suitable for the detection and classification of PQE.

Arun Kumar Puliyadi Kubendran, L. Ashok Kumar

Chapter 4. PMSM Drive Using Predictive Current Control Technique for HVAC Applications

This work deals with Predictive Current Control (PCC) strategy using Field Oriented Control (FOC) for Permanent Magnet Synchronous Machines (PMSM), to improve the performance of the current loop. Based on the research conducted, FOC and PCC have good performance in PMSM drives. Hence a new method based on combined FOC-PCC is adopted in this project work. The main objectives of the FOC-PCC controller are to improve the dynamic characteristics of the drive and to reduce the current ripple. The simulation has been implemented with MATLAB/Simulink for a 750 W PMSM with rated speed of 3000 rpm and rated current of 4.8 A. The sampling time has been set to 20 μs, and the DC link voltage is maintained at 180 V. The reference speed has been set to 3000 rpm and load torque is set to 2 Nm. The stator currents have been limited below its maximum allowable set limit during the dynamic operations such as starting and braking and also during speed reversals with the help of control technique.

L. Ashok Kumar, V. Indragandhi

Chapter 5. Grid-Connected 5 kW Mono-crystalline Solar PV System

Solar power industry in the country is growing rapidly. As of last month (September 2017), the country’s aggregate solar capacity is 16.20 GW. The country increased its solar power production capability to about fourfold from 2650 MW in the month of May 2014 to 12,289 MW in the month of March 2017. The paper aims to offer tools along with strategy so as to ascertain that solar PV power systems are correctly specified and installed, giving an efficient arrangement which works to its rated capacity. The paper addresses solar PV systems set up on rooftop that are interconnected to grid. Statistical data studies reveal that about 10–20% of recently set-up solar PV systems have major drawbacks in installation which has led to performance deterioration, significantly reducing output. Proper sizing and orientation of the solar PV panel to obtain maximum electrical power and energy output are important. Poor public opinions resulting from improperly installed solar PV system will adversely affect the whole solar industry. In this paper, a 5 kW mono-crystalline solar PV system design analysis is carried out. Hardware results of 5 kW mono-crystalline solar PV installation are also presented.

L. Ashok Kumar, Sheeba Babu, V. Indragandhi

Chapter 6. An Add-in Tool for BIM-Based Electrical Load Forecast for Multi-building Microgrid Design

For effective multi-building smart microgrid design and optimization, it is necessary to gather sufficient and accurate computational data that too at the early design stage of projects. The building level electric energy demand forecast is one of the significant steps for suitable generation planning and to formulate strategies for demand response. The adoption of information and communication technologies (ICT) in construction such as building information modelling (BIM) added significant values by improving productivity and by providing digital and object-oriented data-rich realistic integrated models for engineering calculations and coordination. This paper discusses the development of an add-in tool as an Autodesk Revit 2017 plug-in application for the BIM-based load forecast estimation to be used for the electrical design of multi-building microgrids. The calculation proceeds by summing up the individual forecasts of each elementary load components by following a simplified bottom-up approach. The proposed tool is capable of generating realistic electrical load profile for a selected time period on an hourly basis.

Jasim Farooq, Rupendra Kumar Pachauri, R. Sreerama Kumar, Paawan Sharma

Chapter 7. An Investigation on Torque Ripple Minimization of Switched Reluctance Motor Using Different Power Converter Topologies Using Intelligent Techniques

In the modern era, the fast necessities of the electrical drive are the indispensable part of the industries. This paper describes the modelling, analysis and study of various converter topologies fed switched reluctance motor (SRM) along with the speed control which is going to be achieved by the fuzzy logic controller. Now a days SRM is gaining more and more attention in recent high-speed industrial application due to its simplicity and ruggedness. Due to the double saliency nature of SRM drive, the torque pulsations are the vital challenge as the torque pulsations are relatively higher when compared to other conventional machines. The origin behind the high torque ripple in SRM is transferring the torque production from incoming phase to outgoing phase during the phase commutation period. Due to the presence of higher torque ripple, acoustic noise and oscillations are induced in the torque. To minimize the torque ripples, the current modulation technique plays the vital role which excites the stator phase sequentially. This paper proposes the different types of converter topologies such as asymmetric, C-dump and R-dump which are used to excite the stator phases of SRM with low torque ripples. Here the speed control can be achieved by the intelligent technique like fuzzy logic controller. The converters are to be designed using MATLAB/Simulink platform, and the mean values of torque and speed are to be compared and analysed. From the comparison, a proper converter is chosen for the SRM to maintain the minimum torque ripple.

M. Gengaraj, L. Kalaivani, K. Cherma Jeya, P. Eswari Prabha, A. M. Kirthika, M. Vavuniya

Chapter 8. Design of Half-Ring MIMO Antenna to Reduce the Mutual Coupling

We propose a novel, dual-polarized MIMO radiator, which consists of half rings, and included in the ring a square shape at the top side and half circular shape at bottom side operating from 2 to 10 GHz with a microstrip feeding. The suppression of mutual coupling is obtained by maintaining the separation between the patches around 0.25 λ0. The half-ring MIMO antenna is resonating at dual band of frequencies at 7.24 and 8.16 GHz with impedance bandwidths of 430 and 230 MHz, has |S11| < −10 dB in the MIMO range from 2.0 to 10.0 GHz and has a mutual coupling with |S21| < −20 dB. The radiator has a low ECC (envelope correlation coefficient) with values equal to approximately less than 0.025, which will prove that the half-ring MIMO radiator shows better diversity performance. The half-ring MIMO radiator has improved the parameters of reflection coefficient, mutual coupling, realized gain, group delay and real/imaginary impedances.

K. Vasu Babu, B. Anuradha

Chapter 9. Optimal Allocation of Distributed Generation Using Clustered Firefly Algorithm

Integration of distributed generation units (DGs) in distribution systems aims to enhance the system performance. Location and the sizing are the two important factors on the network power loss. This work proposes a clustered firefly algorithm (CFFA) to reduce the distribution system loss, simultaneous optimal placement and sizing of the distributed generation resources in radial distribution systems studied. The simulation is done on IEEE-69 bus network in MATLAB software. The simulated results demonstrate the effectiveness of the proposed clustered firefly algorithm compared with other optimization algorithms.

K. Banumalar, B. V. Manikandan, S. Sundara Mahalingam

Chapter 10. CDM-Based Two-Degree-of-Freedom PID Controller Tuning Rules for Unstable FOPTD Processes

This paper deals with the coefficient diagram method (CDM)-based two-degree-of-freedom proportional integral derivative (CDM-PID) controller tuning rules for unstable first-order plus time delay (UFOPTD) processes. The explicit tuning rules for setting the PID controller parameters are derived using a general UFOPTD transfer function model, the second-order Taylor denominator approximation technique and the pole allocation strategy named CDM. The derived tuning rules are novel, directly relating the controller parameters to the process model parameters. Simulation results indicate that the CDM-PID controller with the proposed tuning rules deliver better performance than other available PID controller tuning methods.

Somasundaram S, Benjanarasuth T

Chapter 11. Real-Time Energy Management System for Solar-Wind-Battery fed Base Transceiver Station

This chapter proposes an intelligent energy management system which integrates solar and wind energy systems with battery backup for making best use of their operating characteristics and obtain better efficiency. Energy management system is programmed for maintaining the energy sustainability in solar–wind renewable energy systems, constant power at point of common coupling, regulating the reference currents based on instantaneous power delivered by the sources and load demand. A controller-based state of charge estimation, charging and discharging of batteries to suitably source or sink the input power according to the load demand is also presented. The simulation of the proposed energy management controller is done using Matlab/Simulink. The simulation results are encouraging in reliability and stability perspective. Experimental results based on solar system with 1.5 kW peak power, wind system with 1.4 kW peak power and 48 V, 200 Ah lead–acid batteries with embedded controller validate the theoretical approach. It is observed that the proposed controller provides constant power at the point of common coupling, thus meets the load demand without interruption.

W. Margaret Amutha, V. Rajini

Chapter 12. IOT-Based Adaptive Protection of Microgrid

Microgrid is a localized power network of connected loads and energy distribution sources that may operate autonomously with respect to the main central grid. Conventional protection schemes are not suitable for grid-connected and islanded modes of operation of microgrid. This chapter proposes an adaptive protection centre (APC) that is capable of real-time monitoring of the microgrid, fault identification and graph algorithms-aided shortest path identification for fault clearance. The Internet of Things (IOT)-based APC is capable of displaying system status and load characteristics in a remote system. The proposed APC is validated and tested for functionality on the hardware prototype of eight-bus microgrid network.

O. V. Gnana Swathika, K. T. M. U. Hemapala

Chapter 13. Performance Comparison Between Sensor and Sensorless Control of Permanent Magnet Synchronous Motor with Wide Speed Range of Operations

Knowledge of rotor speed and rotor position is essential for effective functioning of Field-Oriented Control (FOC) technique. But this requires sensors, and thus employing advanced vector control strategies is challenging in terms of cost and reliability. In this chapter, a Sensorless Field-Oriented Control of Permanent Magnet Synchronous Motor (PMSM) covering wide speed ranges is evaluated. Maximum Torque Per Ampere (MTPA) can enhance the torque output capability, minimize the stator current and thereby copper loss and increase the overall drive efficiency. However, it is not suitable for above base speeds due to limitations on inverter ratings. Domestic appliances such as washing machine require higher speeds during spin dry cycle and usually two to three times of rated speeds. Even traction applications require a wide range of speed control. To achieve speeds greater than rated speeds, Flux Weakening is employed. Model Reference Adaptive System (MRAS) based on stator current controller is used as rotor position estimation algorithm and a comparison is performed between a controller with sensors and a sensorless controller. Results have demonstrated the effectiveness of the sensorless PMSM compared to controller with sensors for both below and above base speed operations.

N. Krishna Kumari, D. Ravi Kumar, K. Renu

Chapter 14. Passive Fault-Tolerant Control Based on Interval Type-2 Fuzzy Controller for Coupled Tank System

In this chapter, a robust controller for a coupled tank-level control is proposed in presence of system component fault. For this purpose, interval type-2 fuzzy logic control approach (IT2FLC) technique is used to design a controller, named passive Fault Tolerant Interval Type-2 Fuzzy Controller (PFTIT2FLC) based on the robust controller to fault tolerant of coupled tank level control system. The proposed control scheme allows avoiding modelling, reducing the rules number of the fuzzy controller. The simulation results show that the PFIT2FLC can provide good tracking performance, even in presence of system component faults.

Himanshukumar R. Patel, Vipul A. Shah

Chapter 15. Enhanced Isolated Boost DC–DC Converter with Reduced Number of Switches

This chapter documents a new two-switch, isolated boost DC–DC converter. At the point when the field effect transistor (FET) is turned on, the component produces a resonant pulse, which is then filtered by the output LC like a traditional switching converter. In a quasi-resonant converter (QRC), the width and amplitude of the pulse are fixed and the converter operates under variable frequency by using a resonant pulse, the switching element will naturally go into a zero-current state. To better comprehend the operation of this circuit, we can simplify our model slightly and break up its operation into six modes of intervals. High boost DC–DC converters are required to convert low voltages into a steady DC bus voltage. For galvanic isolation, segregated topologies are generally used. As a result, the voltage-fed full-bridge (VFFB) DC–DC converters are appropriate for applications with high step-up voltage gain. Regarding the performance of the proposed converter, a 343 V DC output was constructed with a 50 V DC input and the higher efficiency obtained is 99%. To check the execution of the proposed converter, a simulation model has to be developed by means of MATLAB Simulink. The developed simulation model needs to be examined under resistive loading condition.

Anjel J, Gerald Christopher Raj I

Chapter 16. Harmonic Intensity Reduction Technique for Three Phase VSI Drive through Double Randomness

Deterministic Pulse Width Modulation (PWM) methods are popular in industrial applications due to merits. However, the randomized PWM (RPWM) with its cleaner harmonic spectrum is gaining interest for industrial applications required to meet electromagnetic compatibility standards with almost all the earlier merits retained. The proposed Harmonic Intensity Reduction Double Randomness PWM (HIRDRPWM) technique is a hybrid RPWM, which attains the randomness in two ways. The first one is, in the pre-pulse generation stage, through the chaotic frequency generator, which generates a random frequency carrier (triangular) wave. Second randomness is, in post-pulse generation stage, by varying the position of the pulse. The competence in spreading the harmonic power of sinusoidal PWM (SPWM) and the HIRDRPWM is compared using simulation. The distribution of harmonic power in the output voltage of VSI with induction motor load is studied using the MATLAB software. The discussion includes Total Harmonic Distortion (THD) in output line voltage and the Harmonic Spread Factor (HSF).

P. Arulkumar, K. Jaiganesh, N. P. Subramaniam

Chapter 17. PV-Based Multilevel Inverter-Fed Three-Phase Induction Motor with Improved Time and Speed of Response

In the present power scenario, the power quality is most significant in the field of grid-connected and load-connected inverter. A number of techniques were proposed in the field of renewable energy for improving the efficiency and quality of the power. In this article, the time response of the controller was analysed and verified for load-connected multilevel inverter. There are different types of controllers in use, such as Proportional (P), Proportional Integral (PI), Proportional Integral Derivative (PID), Integer Order PID (IOPID), etc. Out of those controllers classic PI controller is most efficient in the speed of response, that’s why the classic PI controller is compared with proposed Fractional Order Proportional Integral Derivative controller (FOPID). Based on the results, FOPID has more speed of response. The outputs of the multilevel inverter are to be improved by minimizing the rise time, settling time and steady-state error of the inverter. The Simulink model was built to control the motor speed and it will going to be applied for grid in the future.

Chandrasekaran S, Durairaj S

Chapter 18. Adaptive Disturbance Observers for Building Thermal Model

Space cooling in buildings is influenced by thermal dynamics, which in turn is affected by ambient conditions, solar radiation, occupancy, stray heating and various other disturbances that are time-varying and nonlinear. This investigation presents an adaptive disturbance observer for estimating the thermal states of the building depending on the disturbance influences. In our approach, the building is modelled as an electrical equivalent circuit and the disturbance influences are modelled as exogenous inputs. Then an adaptive observer is designed for estimating the disturbances and providing accurate state estimates. Further, we also provide the conditions in which the adaptive observer provides an estimate of the disturbance. The proposed approach is illustrated on a test building with an air conditioner controlled using a thermostat. Our studies showed that the proposed observer provided accurate estimation of temperature depending on the disturbances.

Mallikarjun Soudari, Seshadhri Srinivasan, B. Subathra

Chapter 19. HTSA Optimized PID-Based MPPT for Solar PV System

PV system is generally applied with DC–DC converter and a battery. The system is always operated at maximum power point, and its tracking is known as maximum power point tracking (MPPT). Several comparisons have been based on MPPT. In this article, a comparison has been made with conventional perturb and observe (P&O) with intelligent fuzzy logic controller. In this article, the system is designed to operate for an islanding mode without a battery which would reduce the cost of a battery. The model is simulated in MATLAB and the obtained simulated result shows that the performance of intelligent optimized fuzzy logic controller is better than conventional P&O method.

Shashikant, Binod Shaw

Chapter 20. Performance Analysis of UFMC System with Different Prototype Filters for 5G Communication

The Universal Filtered Multi-Carrier (UFMC) system is a promising contender for next-generation 5G wireless communication. UFMC has better sub-carrier separation like Filer Bank Multi-Carrier (FBMC) and provides minimal complexity like Orthogonal Frequency Division Multiplexing (OFDM). Hence, in this chapter, the UFMC system has been proposed with different prototype filters for shaping the waveforms and to increase performance. The performance of the UFMC system is analysed for important parameters such as Power Spectral Density (PSD), Block Error Rate (BLER), Bit Error Rate (BER) and Peak-to-Average Power Ratio (PAPR). The filters considered for the analysis are the Hermite filter, the PHYDIAS filter, the Root Raised Cosine (RRC) filter and the Rectangular filter. The analysis shows that among the four filters, the Hermite filter offers the best performance. This filter provides 72%, 63%, 67% and 29% enhanced performance for the parameters PSD, BLER, BER and PAPR, respectively.

M. Maheswari, N. R. Nagarajan, M. Banupriya

Chapter 21. Fully Convolved Neural Network-Based Retinal Vessel Segmentation with Entropy Loss Function

The eye is the exclusive organ for the sense of sight in humans. Morphological changes in vascular diameter and branching pattern of retinal vessels lead to blindness. Segmentation of retinal vessels is done to analyse these morphological changes in retinal vessels. However, due to the presence of illumination, multiplex distribution of blood vessels, and low contrast between target and background, the task of segmentation of retinal blood vessels is highly challenging. In this chapter to segment retinal blood vessels, we propose a method based on fully convolutional neural networks and pixel classification with cross-entropy function to avoid the class imbalance problem. Our proposed architecture of fully convolutional neural networks combines the output of each stage to learn the hard samples. The cross-entropy loss function is performed to avoid misclassification of vessels.

V. Sathananthavathi, G. Indumathi, A. Swetha Ranjani

Chapter 22. Solar Power Forecasting Using Adaptive Curve-Fitting Algorithm

Electricity is generated from different sources such as thermal, coal, nuclear, solar, wind and so on. The generated electricity is connected to grids for further use. If forecast from renewable energy is available, then the utilization of non-renewable resources could be reduced and so the cost and impact on environment can also be reduced through optimized grid balancing. Solar power is one of the renewable power sources in focus due to the upgradation of photovoltaic technologies and simplified system components. But, the yield out of photovoltaic cells could be strongly influenced by factors such as shadow, cloud, rain, dust, temperature, humidity, panel angle, seasonal effects, panel efficiency and so on. Hence, all these factors need to be considered for forecasting solar power. In this chapter, the ‘Adaptive Curve Fitting’ model is introduced for forecasting solar power, wherein the majority of the algorithm is based on mathematical modelling which considers clear sky reference power, real time power and optionally the weather prediction data.

N. Sampathraja, L. Ashok Kumar, R. Saravana Kumar, I. Made Wartana

Chapter 23. A Review of Electric Vehicle Technologies

Electric vehicles (EVs) have gained remarkable attention due to growing concerns over global warming and the depletion of fossil fuels in the last decade. The propulsion system in EVs comprises electric motors, which are fed by energy storage units through power electronic devices. Due to limitations of conventional energy storage systems in terms of cost, sizing, management, energy and power density, it has become necessary to have an energy generating unit along with the energy storage unit. Effective utilization of energy in EVs can be carried out with the incorporation of advanced power electronics technologies. This chapter reviews the various classifications of EVs, electrical propulsion systems, energy storage systems and energy management systems. This chapter also highlights the various issues to be considered for effective electrification in EVs.

P. Ravi Kumar, C. Gowri Shankar, R. Uthirasamy, V. J. Vijayalakshmi

Chapter 24. Gabor Filter-Based Tonsillitis Analysis Using VHDL

Image analysis finds application in a wide variety of areas, namely tumour detection, security purpose by monitoring the captured images, diagnosis of early-stage diseases in various parts of the body and so on. Image segmentation plays a major role in image processing to improve the form of an input image for its analysis in further steps. Segmentation is a key factor in image analysis to maintain less computational time and to derive proper meaning in the presence of large distractions and noises in the image. The key challenge in image segmentation is to attain faster computations and low cost without affecting the basic features of the image. This chapter presents several of the segmentation methods used in images. They are (1) Region-Based Segmentation, (2) Threshold-Based Segmentation, (3) Cluster-Based Segmentation and (4) Filter-Based Segmentation. We proposed a new method for image segmentation with Gabor filter bank by orientation of filters in all directions from 0° to 360°. In this chapter, the proposed image segmentation with a Gabor filter is applied for tonsillitis disease-affected image and the simulations using MATLAB and Block Memory Generators (BRAM) using Very High Speed Hardware Description Language (VHDL) in the Xilinx tool are shown.

P. Nagabushanam, S. Thomas George, D. S. Shylu, S. Radha

Chapter 25. Incorporation of Modified Second-Order Adaptive Filter in MFGCI for Harmonic Mitigation of Microgrid

Power quality is important in the distributed grid system to supply clean and stable power. A multifunctional grid connected inverter (MFGCI) is intended for effective utilization of Distributed Energy Resources and also to provide continuous supply still in deprived power quality situations. The shunt-series switched (SSS) configuration is used for both current and voltage compensation. Therefore, MFGCI is capable of providing series and parallel tie with the grid and load using bidirectional switches. Conversely, the functionality of SSS-MFGCI is imperfect during voltage compensation. In addition, power quality troubles regarded are partial to harmonics. Thus, a Modified Second-Order Adaptive Notch Filter (SOAF) is implemented for multipurpose control of MFGCI and power quality improvement. The control scheme is multipurpose as SOAF is used for grid synchronization, reference signal generation and total harmonic distortion (THD) reduction. The filter technique is adaptive to circumvent the loss occurred. A modified SOAF is used for its improved disturbance rejection capabilities.

P. C. Keerthiga, G. Gabriel Santhosh Kumar, S. Hemila Haland

Chapter 26. Optimal DAU Placement for Smart Distribution Grid Communication Network

The reliability and stability of communication systems are crucial for utility centres to deliver power to the consumers in an efficient manner. This chapter investigates optimal placement of Data Aggregation Units (DAUs) in smart distribution grids equipped with smart metres and smart automation devices (SADs). The DAUs are used as relay points to transmit the data from smart metres and SADs to the control centre in a cost-efficient manner. The optimization of DAU placement is done based on the iterative K-means clustering method. This work presents an optimal placement of DAUs in a wireless network, which includes smart metres and SADs that help the utilities communicate within themselves with less delay.

S. Premkumar, M. Susithra, V. Saminadan

Chapter 27. Long-Term Forecasting of Hybrid Renewable Energy Potential Using Weibull Distribution Method in Coimbatore

Assessment of hybrid renewable energy is a critical factor for the suitable development of distributed generators (DG). Sizing of DGs needs accurate forecasting. This chapter presents a forecast of wind speed and solar irradiance using the two-parameter Weibull Distribution (WD) method. The dimensionless shape parameter ‘k’ and scale parameter ‘c’ are determined based on hourly global solar irradiance and wind speed from time series data during 2004–13 is used to estimate the Weibull parameters: hourly global solar irradiance and wind speed predicted all around a year. The performance of the Weibull Distribution method is analysed using Mean Absolute Error (MAE) and Mean Squared Error (MSE). The obtained result indicates the EPF method is suitable for prediction of mean hourly solar irradiance and wind speed.

Anuradha J, Soundarrajan A, Rajan Singaravel M M

Chapter 28. Efficient and Improved ANN-Based Voltage-Rise Mitigation Strategy in Distribution Network with Distributed Solar Photovoltaic System

As the cost of photovoltaic (PV) generation systems reduces, more consumers will add grid-tied roof-top PV systems to low-voltage (LV) distribution networks in a widespread manner. Transmission-line impedance and load variations will influence the power transfer capability and voltage profile of the system. This chapter describes the influence of line impedance and load change on the voltage profile in the distribution network. An analysis on the voltage profile due to the variation of load change and line impedance is investigated in a representative LV distribution network and a new control strategy-based on Artificial Neural Network (ANN) is introduced for mitigating the voltage-rise issue in the system. The system is simulated using a MATLAB/SIMULINK platform.

Neenu Thomas, R. Jayabarathi, T. N. P. Nambiar

Chapter 29. Control of Buck Converter by Fuzzy Controller for Wind Energy: Battery System

The scope for wind farm as assessed using GIS platform is found to be more than 2000 GW. At present, India has installed wind farms of approximately 28,000 MW capacity and stands fourth largest in the world. So, the scope of improvement in the figures is immense. As a sustainable power source, wind source has turned out to be a promising one. Much research and development in wind energy conversion systems (WECS) has shown their excellent potential in remote areas located so far from power stations and distribution networks where it is uneconomical to install them.As the output from renewable sources is highly variable, the battery system holds the key to stabilizing it before its integration with load. This chapter presents an introduction to fuzzy logic-controlled buck converters powered by wind energy conversion systems for constant voltage battery charging. It also presents the simulation analysis and comparison between PI controller and fuzzy controller operations of buck converters.

Sheeba Babu, L. Ashok Kumar

Chapter 30. A Survey on Secure Beamforming in MIMO-NOMA-Based Cognitive Radio Network

Cognitive radio network (CRN) and non-orthogonal multiple-access (NOMA) make a significant contribution to fifth-generation (5G) wireless communication systems. But securing multiple-input–multiple-output (MIMO) NOMA beamforming is yet an exclusive way. This chapter reveals the latest survey of the security of MIMO NOMA in 5G engineering, which includes securing by cascaded transmitting downlink zero-forcing-beamforming (ZFBF) technique, general power allocation scheme, applying the NOMA protocol in large-scale networks and applying new bi-directional ZFBF mechanism, cooperative NOMA in both amplify-and-forward (AF) and decode-and-forward (DF) protocols. Also an efficient majorization–minimization (MM) method-based semi-closed form secrecy rate optimization algorithm are reviewed in this chapter.

Thulasimani Lakshmanan, Hyils Sharon Magdalene Antony

Chapter 31. Hybrid Optimization of Cuckoo Search and Differential Evolution Algorithm for Privacy-Preserving Data Mining

Data mining applies data analysis techniques to find patterns as well as relations in information to make good decisions. In today’s digital world, preserving data privacy is a challenging task due to business enterprise applications’ leverage to modern technologies. The k-anonymity model is used to protect each single record from being identified by making all records indistinguishable from k-1 other records. Suppression and generalization methods are used to implement k-anonymity. In this chapter, generalization method is used to implement k-anonymity. To improve the accuracy of a classification algorithm, hybrid optimization of Cuckoo Search (CS) with Differential Evolution (DE) search technique is used to select optimized generalized feature set. The classification accuracy is evaluated using public dataset with and without k-anonymity.

J. Sudeeptha, C. Nalini

Chapter 32. Using Sliding Window Algorithm for Rainfall Forecasting

Rainfall forecasting has been an onerous task to deal with. However, it is a part and parcel to sustain our life since it affects not only the agriculture growth but also the farming community. Rainfall prediction will definitely pose a great challenge but for meticulous planning and management of water resources. Therefore, this chapter presents an approach for rainfall prediction using Sliding Window concept with Jaccard distance metric measure. The Sliding Window Algorithm watches the information during a similar period in an earlier year and predicts precipitation in the next year. Using Sliding Window Algorithm, the precipitation expectation test was tested for Tirunelveli District, Tamil Nadu, India, using the rainfall data for a 10-year period.

M. Vijaya Chitra, Grasha Jacob

Chapter 33. Air Pollution-Level Estimation in Smart Cities Using Machine Learning Algorithms

Air pollution is a serious issue that has been harming and hurting the people in India. The air is contaminated owing to industrial plants and manufacturing activities, combustion from fossil fuels, farming chemicals and household products and natural events like volcanic eruptions, forest fires and gaseous releases from decaying plants and animals. Air pollution not only harms the comfort and health of both humans and animals but also destroys the life of the plants. Air pollution is otherwise called as environmental pollution that causes serious problems confronting humanity and other life forms on planet Earth today. In this work, K-nearest neighbour method is used to evaluate the position of air pollution at several places in Chennai City. Random Forest and Support Vector Machine algorithms evaluate the efficiency of the proposed model, thereby categorising the data into six classes of pollution levels.

M. Nelgadevi, Grasha Jacob

Chapter 34. Implicit Continuous User Authentication Using Swipe Actions on Mobile Touch Screen with ANN Classifier

Smart phones became the most trusted companion of men and women. Today a smart phone has all the important data within the phone memory or in the cloud which is directly accessible by the device. The sensitivity of these data varies from person to person. Commonly, the security of a smart phone lies with normal entry point authentication methods such as PIN and graphical passwords. These methods can be breached by shoulder surfing, smudge attack, etc. But the main point is that the device is not checking the genuineness of the user after the entry point authentication. Most users prefer simple GUI passwords or PIN or no password at all (Bonneau J, The science of guessing: analyzing an anonymized corpus of 70 million passwords. In: 2012 IEEE symposium on security and privacy, San Francisco, CA, pp 538–552, 2012). So a smart phone after the primary authentication can become a threat to losing sensitive and private data. In this work, we present an implicit continuous active authentication mechanism that will check the genuineness of the current user without any direct input. We would be using the touch screen swipe patterns that are being generated when reading a page or viewing the images in gallery of the phone. We are using the artificial neural network to recognize the genuine user. Result shows the proposed mechanism has an accuracy of 93.9% and an EER of 7%, and it is not a burden to the user as he is not supposed to make any deliberate inputs, the data generated from the normal usage is taken for authentication.

Christy James Jose, M. S. Rajasree

Chapter 35. A Review on Graph Analytics-Based Approaches in Protein-Protein Interaction Network

Essential proteins play a vital role in the biological and cellular activity of a living organism. Identification of essential proteins is crucial for understanding the cellular life mechanisms for medical treatments and disease diagnosis. The existing computational measures are primarily based on identifying dense sub-graphs from the protein interaction network. In this research paper, the existing computational, graph theoretic approaches are reviewed and a novel research direction to find essential proteins is proposed.

D. Narmadha, A. Pravin, G. Naveen Sundar, Premnath Dhanaraj

Chapter 36. A Survey on Emotion Detection Using EEG Signals

Emotions play a huge role in the social interactions between people which makes it important to study its working to make intelligent humanoids that can socialize on a higher dimension than was deemed possible in the last century. This paper gives a survey on the various approaches used to detect emotions and explains the role of the brain in generating human emotions. A review is also made on the available classifiers and latest trends used in analyzing the EEG signals.

Oshin R. Jacob, G. Naveen Sundar

Chapter 37. A Smart Agricultural Model Using IoT, Mobile, and Cloud-Based Predictive Data Analytics

In recent times, the amount of data generated by the IoT devices is very huge and the traditional databases do not have enough storage space. So the need for cloud storage becomes essential. Data mining techniques are used to analyze this big amount of data available in cloud. The study of smart agriculture system has cloud-based data analytics with IoT as a major role. The role of Information and Communication Technologies in the field of smart agriculture model is very important to extract the information from the field. In this paper, the IoT device records the data from the agriculture field and stored in the cloud database. Data analysis is done on the data available in cloud, and based on the data mining technique used, the prediction is performed. The predicted information is sent to the farmer through a mobile phone application. The main aim is to increase the production and reduce the cost of the products based on the predicted information.

P. Anand Prabu, L. S. Jayashree

Chapter 38. Machine Translation Using Deep Learning: A Comparison

In recent days, machine translation is rapidly evolving. Today one can find several machine translation systems that provide reasonable translations, although they are not perfect. The main objective of machine translation is to provide interaction among the people speaking two different languages. Machine translation, being an important task of natural language processing, leads to the development of different approaches, namely, rule-based machine translation, statistical machine translation, and neural machine translation for the translation process. The recently proposed method is the neural machine translation which improves the quality of translation between natural languages through neural networks. Neural machine translation led to remarkable improvements in the translation process by retaining the contextual information. End-to-end neural machine translation uses RNN Encoder-Decoder mechanism to train the neural translation model with bilingual corpora which is bilingual parallel sentence pairs, an important resource of machine translation. NMT has a reasonable BLEU score which is the evaluation metrics for machine translation. In this paper, we present a survey on the different kinds of machine translation approaches with their strengths and limitations and the various evaluation metrics to measure the accuracy of the translation.

S. Swathi, L. S. Jayashree

Chapter 39. Societal Impact of Framework for Energy-Efficient Clustering Algorithms in Mobile Wireless Sensor Networks

Mobile Wireless Sensor Networks (MWSNs) are capable of sensing various types of events and change their position frequently in a specific sensing area. Architecture, energy, mobility, degree, distance, topology, localization, and data collection are the key factors in designing an energy-efficient MWSN. The applications of MWSNs can be widely divided into single-based, grid-based, and high density-based area applications. FEECA (Framework for Energy-Efficient Clustering Algorithms) is proposed to serve the various types of MWSN applications with reduced energy consumption and increased network lifetime. This paper deals with the societal impact of FEECA in various applications.

K. Juliet Catherine Angel, E. George Dharma Prakash Raj

Chapter 40. Energy Demand Prediction Using Linear Regression

Big Data analytics is the latest emerging technology that requires deep knowledge in business intelligence, machine learning, and statistical methods and in deep learning. It focuses on the application of data analytics for energy demand management using real-time data. The data is then analyzed for clustering, demand forecasting, pricing, and energy generation optimization. It represents a method to predict energy usage, based on real-time data obtained from TANGEDCO-CBE, using the linear regression model (LR). The final linear regression models developed were based on daily sustained demand and consumption by comparing actual and predicted energy usage models can predict with acceptable errors. Normally the energy requirement and industrial demands are high; hence the application of these energy Big Data analyses significantly improves efficiency and provides new business opportunities.

T. Manojpraphakar, Soundarrajan A

Chapter 41. Risk Prediction Analysis of Cardiovascular Disease Using Supervised Machine Learning Techniques

The best thing to avoid strategic human death rates due to curable diseases is to detect them early and prevent their onset. Presently, in our society, large numbers of death rates are due to cardiovascular disease (CVD). Hence early detection of CVD is critical even though many more practices exist for earlier prediction of risk. One approach for early disease risk prediction is the use of risk prediction models developed using machine learning techniques. These models will provide clinicians to treat heart disease of the patient in a better way. Consequently in this chapter, classification mechanisms have been applied to predict the status of the disease. The machine learning algorithms involved in the prediction of CVD are EDC-AIRS, Decision Tree, and SVM. The heart disease dataset from UCI repository has been used in this study. The predictions are denoted by means of accuracy, whereas the performance measures have been calculated in terms of sensitivity, specificity, and F-measure. Results indicate that the prediction model developed using the SVM algorithm is capable of achieving high sensitivity, specificity, balanced accuracy, and F-measure. Further, these models can be integrated into a computer-aided screening tool which clinicians can use to predict the risk status of CVD after performing the necessary clinical assessments.

A. Ishwarya, S. K. Jayanthi

Chapter 42. Safest Secure and Consistent Data Services in the Storage of Cloud Computing

Cloud computing is the greatest learning in the computing field and a dreamed vision of computing as a utility so to enjoy the on-demand high-quality applications. Cloud security is the critical factor that places an imperative role in maintaining the secure and reliable data services. In large-range cloud computing, a large pool of erasable, usable, and accessible virtualized resources are used as hardware development platforms and/or sources. These resources can be vigorously reconfigured to adjust a variable load allowing also for optimum resource utilization. The pool of resource is typically exploited by a peer-to-peer use model in which guarantees are presented by the infrastructure provided by means of customized service-level architecture (SLA).The hierarchical structure has been proven effective for solving data storage issues as well as data integrity by giving data protection during the full life span. Cloud computing is related to numerous technologies, and the convergence of diverse technologies has emerged to be called cloud computing. Storage in the cloud provides attractive cost and high-quality applications on large data storage. Security offerings and capability continue to increase and vary between cloud providers. Cloud offers greater convenience to users toward data because they will not bother about the direct hardware management. For security issues, a secret key is generated. Key consideration is to efficiently detect any unauthorized data corruption and modification which arises due to byzantine failures. Cloud service providers (CSP) are separate administrative entities, where data outsourcing is actually relinquishing user’s ultimate control over the fate of their data. As an outcome, the accuracy of the data in the cloud is being set at a high risk. In distributed cloud servers, all these inconsistencies are detected and data is guaranteed. The main proposed objective of this chapter is to develop an auditing mechanism with a homomorphic token key for security purposes. By using this secret token, we will easily be able to locate errors and also the root cause of the error. By the error recovery algorithm, we recover these corrupted files and locate the error.

Geethu Mary George, L. S. Jayashree

Chapter 43. Agile Supply Chain Management Enabled by the Internet of Things and Microservices

The business networks of modern world span across the globe, and their requirements for supply chain management and logistics have grown extensively. The field is continuously innovating to make the supply chain more predictable, to optimize the logistics, and to be cost-efficient. This paper investigates the impact of using the Internet of Things in supply chain management in the first part. The software engineering methodologies are rapidly evolving to meet the current business requirements of the modern world, and companies are moving toward more agile methodologies. Microservices architecture solves numerous problems in the modern software development process. The impact of following microservices architecture is analyzed in the second part of the paper. Finally, all the dots are connected by discussing the development of IoT-enabled supply chain solutions with microservices architecture. While microservices, IoT, and supply chain management are evolving in their own aspects, this paper is an attempt to bring all the three together to achieve more agile supply chains to satisfy the growing business needs.

G. Selvakumar, L. S. Jayashree

Chapter 44. Production and Characterization of Bio-Fertilizers from Tree Leaves Utilizing an Automated Hot Composting Chamber with Cyber-Physical Systems

Composting is a reliable process to transform waste dry leaves into superior quality compost. In this research, we proposed a rapid composting technique for converting powdered dry leaves to mature compost in an automated rotary drum composter and for monitoring the physicochemical properties of the pile by using cyber-physical systems. Here, the solar energy-based equipments were utilized for grinding dry leaves and energizing a heating coil inside the composter. The optimal temperature of the composter had been optimized and maintained at 50–55 °C by using closed-loop heating coils and monitored by using temperature sensors. During composting, the moisture content (MC) decreased with respect to the duration. On the contrary, the pH and electrical conductivity (EC) gradually increased as the dried leaves are being converted to mature compost. The aim of this research is to automate the composter by using sensing devices and monitor the physicochemical characteristics in terms of pH, electrical conductivity, moisture content, and temperature. Here, we anticipated a cost-effective, less maintenance, and eco-friendly quick-composting technique for obtaining good quality of compost from the leaves within 30 days.

Mahendran Rajagopalan, Vijayakumar Arumugam, Uma Dharmaligam, Kavinilavu Anbalagan, Anupriya Chandrasekaran

Chapter 45. Spectrum Sensing Based on Cascaded Approach for Cognitive Radios

The swift growth in radio communication technology has drawn to scarcity in wireless spectrum. Literature points out that licensed spectrums allotted by regulatory agencies are underutilized in the allocated band of spectrum. Cognitive radio networks appear to be a promising solution to address the bandwidth scarcity and demands of wireless spectrum. Cognitive radios are confronted to maximally utilize the spectrum through sharing of spectrum with the licensed primary users (PU). The efficiency of cognitive radios mainly depends on the efficiency of the spectrum sensing plane, in which better spectrum utilizations are exploited. A hybrid approach of merging energy detection (ED)-based channel sensing and cyclic prefix autocorrelation detection (CPAD) techniques has cascaded in the way to boost the probability of detection, which is proposed in this chapter. Energy detection techniques implicate no prior knowledge of PU signals, less computational complexity, and low energy consumption but hold uncertainties at low SNRs (at −20 dB to 10 dB). The next mentioned CPAD technique is more robust at less decibels, but it requires a large number of samples yielding complexity and increase in sensing time. Consequently, ED and CPAD techniques, with an influence on the benefits of each technique, have been designed as cascading and implemented using the Universal Software Radio Peripheral (USRP) tool. On comparison, the probability of the detection of ED and CPAD at SNR ranging from −20 dB to 5 dB is around 0.3–0.9 and 0.6–0.9, respectively. The cascaded design has the probability of detection in the bound 0.7–1 for the same specified SNR. Spectrum sensing based on cascading a couple of detectors outperforms in detection probability compared to a single detector.

N. Iswarya, L. S. Jayashree

Chapter 46. Remote Process Monitoring and Control Through IIoT

Internet functionalities and data-based services are introduced in industrial manufacturing processes to innovate the process control based on the Industry 4.0 standards. In this regard, the vision of the IoT extends to closed-loop control. Sensors connect through control algorithms to actuators, with communication over the cloud. Also, the sensor data accumulated in cloud are utilized for process data analytics. These analytics perform predictive control actions by pattern identification, anomaly detection, and outcome prediction from the raw data, as well to trigger control actions through the application of rules. To demonstrate closed-loop control through the IoT, the level of a nonlinear continuous stirred tank reactor plant is continuously monitored and controlled wirelessly using a Wi-Fi-enabled microcontroller CC3200 (Texas Instruments) through an IoT analytics cloud platform. At every time point, the measured values are automatically updated in the cloud. The performance of the IoT-enabled control system is validated by changing various set points. The main implication of this work is to securely and remotely monitor and control various parameters in industry, universally through Internet affiliation.

Swetha R Kumar, Sangavarthini C.S., L Ashok Kumar

Chapter 47. Case-Based Reasoning (CBR)-Based Smart Indoor Navigation

In order to improve the efficiency of an urban infrastructure, the concept of smart cities can be used, which is an equal balance between communication technologies and physical devices that connect to them. The recent changes in technology and economy have garnered a significant level of attention in the field of smart cities. The evolution of a city, in terms of both the city infrastructure and the community, can be monitored directly through the help of this technology. One such domain under the smart city infrastructure is the medical domain. On a global scale, hospitals need to take care of factors like pressure with respect to cost and reimbursement, as this can happen when they serve a population with illness. Hence, they need to find out ways to improve their efficiency. By combining the concept of smart city with hospitals, we get smart hospitals. Smart hospitals are being developed with the intention of providing a better value-added service for the common people. It also helps in redesigning and radically thinking about new processes. All of these are enabled by an interconnection of a complete network infrastructure. Further analysis of the smartness in hospitals would lead us to localization. Localization has to heavily depend on sensing devices, such as beacons, RFID tags, etc., despite the use of GPS services. Taking a typical hospital environment into account here, a navigation path should not only provide the shortest path but also keep track of the overall paths the patient has traversed across the hospital. This will help in analyzing the paths of the specific patient and give the respective recommendation measures that need to be taken in order to complete the course of action (in this case, diagnosis and treatment). The motive is a smart indoor navigation system, which learns the user’s behavior through previous sensing data.

G. R. Karpagam, K. Eshwar, K. Karthikeyan, M. Syed Hameed

Chapter 48. A Survey on Medical Image Registration Using Deep Learning Techniques

Image registration is one of the most significant and useful approaches in diagnosing disease by providing complementary information from different medical images. Image registration is a process of overlaying two or more images into a single integrated image. This process is widely used in medical imaging analysis to overlay images obtained from different devices at different time. Traditional methods to geometrically align images are time-consuming, while deep learning techniques are less time-consuming. In recent years, deep learning is a growing technology and has gained many breakthroughs in various image processing problems such as classification, reconstruction, and registration. In particular, convolutional neural networks (CNNs) is one of the most powerful tools in computer vision task. Recently, deep learning techniques are being developed for medical image registration, and image fusion is clearly evidenced from high-quality research. The intention of this survey is to provide perspective about the recent development of registration techniques using machine learning and deep learning techniques.

M. C. Shunmuga Priya, L. S. Jayashree

Chapter 49. Agent-Based Temperature Monitoring System

The field of IoT is growing day by day; this growth will lead to wide applications and devices to be interconnected with the help of the Internet. However, there are still many hurdles which have to be rectified before all these will be possible. In this chapter, we have addressed the issue of organising and adapting to the environment where the IoT system is present. To achieve this, we have used the concept of multi-agent systems.

S. Jaswanth, L. S. Jayashree

Chapter 50. Classification of Phonemes Using EEG

Artificial speech synthesis can be done using electroencephalography (EEG) and electrocorticography (ECoG) for the brain–computer interface (BCI). This paper focuses on using EEG to classify phonological categories. Although literature is available on the identification and classification of phoneme information in the electroencephalography signals, the classification accuracy of some phonological categories is high, while that of others is too low. Thus, this chapter focuses on identifying the correlation between imagined EEG and audio signals to select the appropriate EEG features. It also identifies the EEG channels that are best suited for imagined speech. Once features are selected, phonemes are classified as vowels and consonants using a support vector machine. Experimental results suggest good accuracy when using 49 features that correlated with audio signals.

R. Aiswarya Priyanka, G. Sudha Sadasivam

Chapter 51. Attribute Table-Based Multipath Routing Protocol to Improve Network Lifetime in Multi-hop WSN

Unbalanced energy consumption is an inherent problem in WSNs characterized by multi-hop routing and many-to-one traffic pattern. This uneven energy dissipation can significantly reduce network lifetime. In this paper, we proposed an attribute table-based energy-efficient and QoS multipath routing protocol to improve network performance in the multi-hop network. The attribute table is prepared for each transaction, and the energy level of the node is monitored for every transmission. The optimized pair shortest path algorithm is used to find the shortest path between all pairs of vertexes and reduce the duration of data transmission. After every transmission node, the energy level is noticed by the attribute table for the next process. This efficiently transmits the data with security and minimizes the packet losses.

B. Sherin, M. Senthil Vadivu, A. Ayub Khan

Chapter 52. Application of Subjective and Objective Integrated Weightage (SOIW) Method for Decision-Making (MADM) in Distribution System

The term smart grid (SG) has been used by many government bodies and researchers, which refers to the new trend in the power industry to modernize and automate the existing power system. SG must utilize the assets optimally by making use of information, like equipment capacity; voltage drop; radial network structure; minimizing investment, operating costs, and energy loss; and reliability indices. One way to achieve this is to reroute or reconfigure the distribution system. The distribution system is reconfigured to choose a switching combination of branches of the system that optimize certain performance parameters of power supply while satisfying some specified constraints. In this chapter, the subjective and objective integrated weightage (SOIW) multiple attribute decision-making (MADM) method is proposed for finding the compromised best configuration and comparing it with other methods such as WSM, WPM, and TOPSIS. An example of the distribution system is presented in this chapter to demonstrate the validity and effectiveness of the method.

Sachin Gorakh Kamble, Kinhal Vadirajacharya, Udaykumar Vasudeo Patil

Chapter 53. Visual Importance Identification of Natural Images Using Location-Based Feature Selection Saliency Map

The proposed saliency map is called location-based feature selection saliency map (LBFSM). The research introduced a new method for identifying the image visual objects and region of unimportance. The saliency map uses Fourier transformation function for feature selection. The proposed method was applied over created natural images collected from different parts. The method’s efficiency was calculated based on objective and subjective quality assessment matrices such as processing time, precision and recall values and receiver operator characteristic (ROC) values. The quality assessment study showed the proposed saliency method efficiency in finding the local and global features from the image. The performance of the state-of-the-art saliency calculation method was experimented on the same natural image dataset. Five different saliency maps and their performance were compared and evaluated based on subjective and objective measures. Nine hundred (CRIST900) natural images were experimented using MATLAB R2015a, and their quality assessment was done using the same software platform. This research gives a conclusion that the result of processing time, receiver operator characteristic (ROC) curve, precision, and recall values provide good performance compared to the state-of-the art saliency map calculation methods.

Malayil Abhayadev, T Santha

Chapter 54. Missing Values and Class Prediction Based on Mutual Information and Supervised Similarity

In recent times, the uses of data mining techniques have increased tremendously due to the increase in a large amount of data. Data mining techniques have been used for many research purposes. But mostly, they all face a single unique problem and that is the missing values of data. During research, large datasets are taken as processed for experimentation of algorithms, and if there is a missing value, these instances are either ignored or any default values are replaced during pre-processing of data. But this way is not correct. In this chapter, a novel prediction technique is proposed that can be used to predict the missing values of a given dataset or a dataset sample by calculating the mutual information, supervised similarity, and cosine similarity. The proposed approach calculated the missing values accurately, and this is experimented using a sample cancer dataset with missing gene values. The proposed prediction technique can also be used to predict class values of new instances of dataset. The experimentation shows that the predicted missing values and class labels coincide with the existing gene subsets and are said to be reliable and accurate.

Nagalakshmi K., S. Suriya

Chapter 55. Fake Product Review Detection and Removal Using Opinion Mining Through Machine Learning

Machine learning is one of the growing trends in artificial intelligence and deep learning scenarios where the machine learns to acquire data from previous cases and implements the data for future prediction and analysis. The objective of this chapter is the detection and removal of fake reviews in online reviews. Majority of online buyers rely on product reviews before making purchase decision of their chosen brand; however, fake reviews pose a continuous threat to the integrity of the product, portals and the easy-to-find reviews on specific products. This chapter aims to develop a system to identify and remove fake reviews with the view of protecting the interests of customers, products and e-commerce portals. Thus, in this proposal, the primary goal is detecting unfair reviews on Amazon reviews through Sentiment Analysis using supervised learning techniques in an E-commerce environment. Sentiment classification techniques are used against a dataset (Amazon) of consumer reviews for smartphone products. Precisely, we use three different algorithms, logical regression algorithm, linear regression algorithm and neural networks (CNN and RNN models), of supervised machine learning technique to find similarities in the review dataset and group similar datasets together to explore unfair and fair positive and negative reviews, which involves screening, collaborative filtering, and removing with an optimal accuracy rate. The core focus or the highlight of this chapter is to explore an algorithm using deep learning that ensures optimal accuracy in the identification of fake reviews.

Minu Susan Jacob, Selvi Rajendran, V. Michael Mario, Kavali Tejasri Sai, D. Logesh

Chapter 56. Ask Less: Scale Market Research Without Annoying Your Customers

Market research is generally performed by surveying a representative sample of customers with questions that include contexts such as psychographics, demographics, attitude, and product preferences. Survey responses are used to segment the customers into various groups that are useful for targeted marketing and communication. Reducing the number of questions asked to the customer has utility for businesses to scale the market research to a large number of customers. In this work, we model this task using Bayesian networks. We demonstrate the effectiveness of our approach using an example market segmentation of broadband customers.

Venkatesh Umaashankar, S. Girish Shanmugam

Chapter 57. Preferential Resource Selection and Scheduling of Cloud Resources Pivot on Value of Information

The selection of resources and scheduling in the cloud are crucial due to the involvement of various features. Scheduling an appropriate resource onto the cloud is influenced by quality of service parameters. Providing a relevant resource to the user consists mainly of three steps: (1) finding the feasible set of resources, (2) selecting the most appropriate resource from the practical set of resources, and (3) scheduling the resource to the relevant processor. Selecting a relevant resource is modeled as a multi-criteria decision-making problem. Factors like availability, trust, cost, responsiveness, reliability, and capability have effects on the resource selection. In this chapter, an efficient workflow has been put into suggestion in consideration to make a selection of the most significant resource using PROMETHEE methodology, and scheduling is performed using a non-pre-emptive priority algorithm. The choice of the optimal resource is done pivoted on the value of information that is requested by the users for all the influencing factors. The outcome of the simulation proves that the suggested workflow decreases the response time, makespan, and cost, which also maximizes the quantity of resources utilized before the deadline.

Renu Suresh Ganvir, Salaja Silas, Elijah Blessing Rajsingh

Chapter 58. A Survey on Supervised and Unsupervised Learning Techniques

Supervised learning is the popular version of machine learning. It trains the system in the training phase by labeling each of its input with its desired output value. Unsupervised learning is another popular version of machine learning which generates inferences without the concept of labels. The most common supervised learning methods are linear regression, support vector machine, random forest, naïve Bayes, etc. The most common unsupervised learning methods are cluster analysis, K-means, Apriori algorithm, etc. This survey paper gives an overview of supervised algorithms, namely, support vector machine, decision tree, naïve Bayes, KNN, and linear regression, and an overview of unsupervised algorithms, namely, K-means, agglomerative divisive, and neural networks.

K. Sindhu Meena, S. Suriya

Chapter 59. Performance Study of IPv6/IPv4 MANET (64MANET) Architecture

Today, the Internet usage has increased due to its efficiency and performance. IPv4 and IPv6 are the most common protocols used for providing Internet communication among the mobile users. There is a big consideration regarding the mobile users’ need to access the Internet by different versions of the protocol in MANET. So far, no authors have proposed to design an architecture for intermobility and intertransaction for the transition communication in MANET. For that, 64MANET architecture is proposed to allow the mobile node to roam from one version to another version of the network along with the features of addressing mobility and transition mechanisms of MANET nodes. The performance of 64MANET is evaluated by different performance evaluation metrics. The results show that 64MANET is suitable for intermobility and interoperability between IPv4 mobile node and IPv6 mobile node in MANET communication.

S. Manimozhi, J. Gnana Jayanthi

Chapter 60. Internet of Things: A Technical Perspective Survey

The Internet of Things (IoT), one of the recent emerging technologies in ICT, has a great potential impact on how we live and particularly in every walk of our life. This paper overviews the Internet of Things, highlights the enabling technologies for IoT, outlines the architectural building blocks, and presents the research challenges and issues. The paper is aimed to give an eye opener and to direct the researchers toward IoT in order to understand the concepts and be aware of the burgeoning technology.

S. Margaret Amala, J. Gnana Jayanthi

Chapter 61. Analysis on DGHV and NTRU Fully Homomorphic Encryption Schemes

Homomorphic encryption (HE) is an emerging scheme that allows computation over encrypted data. The standard encryption algorithms like RSA, Elgamal, etc. help in protecting confidential data from attackers rather than performing computation over encrypted data. Fully homomorphic encryption (FHE) permits computation to perform upon encrypted data unlimitedly in server side than in computational node. In this paper, the basic DGHV FHE scheme and NTRU FHE scheme are analyzed to preserve the security and privacy of the data. DGHV performs computing over real integers, while NTRU in a truncated polynomial ring. A detailed investigation of both the schemes is based on the storage and noise reduction that best suits for a real-world application.

B. Santhiya, K. Anitha Kumari

Chapter 62. Automated Image Captioning for Flickr8K Dataset

Automated, accurate image captioning is currently a hot topic in the field of deep learning. The model must have the capability to generate human-readable sentences for regions in the image. The model must understand the image to find the words that string together to be comprehensive. To achieve this, in this research work, Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) are used on Flickr8K dataset. To identify the regions in the image and to recognize the objects in the regions, an advanced region-based CNN (RCNN) methodology has been used. To generate the caption that is most relevant to the image, RNN is used in this paper. Bilingual evaluation understudy (BLEU) score is considered as the evaluation parameter.

K. Anitha Kumari, C. Mouneeshwari, R. B. Udhaya, R. Jasmitha

Chapter 63. RAkEL Algorithm and Mahalanobis Distance-Based Intrusion Detection System Against Network Intrusions

With a quick increment in the volume of information in everyday life, there is dependably a requirement for an intrusion detection framework which distinguishes and identifies the attacks at a quicker pace. Any action which abuses the approach of the security premises is characterized as an intrusion. Intrusion detection system (IDS) is an hardware that cocurrences with intrusions produced by an alternate host network frameworks and system sources, then looking at the sign of security issues. IDS is also used to recognize unapproved usage of PC, which ought to interface the gaps in against antivirus and firewall. A general issue in the current IDS is the high false-positives and low-detection rate. This chapter talks about the essential intrusion detection procedures by means of live capturing of network packets. The proposed system uses Mahalanobis distance methodology in best attribute selection and exhaustive search feature selection methods for feature ranking and removal of features for choosing the superlative possible combinations of features from the feature set obtained from the network packets. The RAndom k-labELsets (RAkEL) multi-label ensemble learning algorithm in combination with machine learning algorithms, like J48, support vector machine (SVM) and Naïve Bayes (NB), are utilized to build up the proposed IDS by classifying different network intrusions with higher detection rate and lower false-positive rate.

R. Padmashani, M. Nivaashini, R. Vidhyapriya

Chapter 64. Vaguely Node Classification Scheme for Wireless Networks to Design an Intrusion Detection System

In recent time, wireless networks that are deployed in various applications are facing many problems due to the presence of malicious nodes in the network. In wireless networks, the nodes may exhibit their mal-behavior temporarily or permanently under adverse environmental conditions. Modification of message and dropping the packets are the two main attacks considered in this paper. In these attacks, malicious nodes may behave like a normal node, but involve in unwanted activities. To identify such malicious nodes, a novel vaguely node classification scheme is proposed in this chapter. It may help to derive the rule sets for identifying the malicious nodes in the network which may perform the unknown attack. In turn, it will be more helpful to design an effective Intrusion Detection System for the wireless network.

S. Latha, V. Sinthu Janita Prakash

Chapter 65. Dynamic Traffic Light Scheduling for Emergency Vehicles Using Fog Computing

This chapter presents a dynamic traffic light scheduling system, which schedules the time of the traffic light controller on each side. Each emergency vehicle is assigned with priority and it has a unique identity (ID) by default. The unique ID and the priority are used to process the request from each vehicle. In the case of priority, conflict distance is considered for scheduling. Various technologies like ZigBee, LoRaWAN are used to provide cost-effective solutions. ZigBee is one of the high-level communication protocol in IEEE 802.15.4. It works at low power, and it has low data rate. ZigBee has a range of 10–100 meters with line of sight communication. The rate of data transmission in ZigBee is 250 kb/s.

S. Sarathambekai, T. Vairam, A. Dharani

Chapter 66. Cloud Database – A Technical Review

Recently, cloud computing technology has attracted several business organizations due to its wide delivery of computing resources as services. As a result, a service computing has emerged from the technology foundation perspectives such as Service-Oriented Architecture (SOA) and virtualizations of software and hardware. The increasing needs have database, information and expert knowledge systems have included vital role in cloud computing. Several issues and challenges are still to be focused and addressed with solution in cloud. Henceforth, this chapter is associated with developers, scientific persons, and users by highlighting and presenting the challenges and issues of cloud database.

S. Sakthivel, J. Gnana Jayanthi

Chapter 67. Projection of Population and Prediction of Food Demand Through Mining and Forecasting Techniques

Food, clothing and shelter are the basic needs of man. The most essential among the needs is food. There is always a wide gap between supply and demand because of the changes in food preference. The change in food preference is the major factor in prediction of food demand. The proposed method uses the second-order Taylor series for the projection of population; having the estimated population as input as the food demand based on the change in food requisite is anticipated. The implementation is carried out through Java. The population and food demand of the continental U.S. are projected by the proposed method. The food demand prediction through the proposed method is similar to the actual demand with deviation close to 0.1%.

J. Antonita Shilpa, V. Bhanumathi

Chapter 68. Detection of Hairline Fracture Foot Using Canny Operator and Wavelet Packet Transform

Bone fracture detection is the most common and usual technique in the medical imaging field. This detection is nowadays easy, since there exist various methodologies for diagnoses. However, detecting open fractures or severe fractures are easy by these detection techniques. While taking the hairline fractures, the case becomes complicated. It is not always possible to detect hairline fractures through bare eye. This is also a challenge for medical imaging. In this study, detection of hairline bone fracture is a problem to be solved via implementing Canny Operator and Wavelet Packet Transform. The preprocessed digital X-ray input image is transformed, and the fracture is detected using Canny Operator. Using Canny edge detection and Wavelet Packet Transform, hairline fractures can be detected in the earlier stage itself. The results are encouraging the effectiveness of Canny edge detection and four-level Wavelet Packet Decomposition is an image enhancer, which is proved in the study.

D. S. Karthika, K. S. Biju, G. H. Silpa, C. Girish Kumar

Chapter 69. Image Encryption-Then-Compression System for Secure Transmission via Hybrid Henon Chaotic Map

In recent years, the environments like military, government, medical field, cloud computing, and social networks deal with a large number of confidential images transmitted over the Internet. Therefore, it is very important to protect the image from unauthorized access during the transmission in an open network. Encryption is the most convenient technique to guarantee the security of images over public networks. To maximize the network utilization, the compression technique is used to reduce the size of the image by the channel provider who is having plenty of computational resources. This paper proposes an image Encryption-Then-Compression (ETC) system. Hybrid Chaotic method is used to encrypt the image, where Arnold map is used for confusion and Henon map is used for diffusion. Asymmetric Numerical Method (ANM) is used to compress the encrypted image. The experiment result shows that the proposed system is better in terms of compression performance, security, and computation time.

P. Sridevi, J. Suguna

Chapter 70. Analysis of Primary Emulsion Attack in Cognitive Radio Using Distributed On-Demand Routing Protocol

The aim of this chapter is to design a novel framework for cognitive radio to overcome the security-based challenge by considering authentication and confidentiality. Particularly, the chapter focuses on the primary emulation attack, as it gives the authentication for the unlicensed user to use the unused the spectrum. These unlicensed users are considered as the secondary users and to authorize to use the spectrum only for the required period without compromising the security of the primary user. The distributed on-demand routing protocol is used in cognitive radio, and hence it can be used for the group of users sharing the same spectrum. RSA with the distributed on-demand routing protocol yields a secure key for sharing that particular session within the users. A comparison between the classical protocols for generating the secret key with Diffie–Hellman algorithm and other protocols is also done in this work by analyzing their vulnerabilities.

Neelaveni Rangaraj, Sridevi Balu

Chapter 71. Heart Disease Prediction Using Retinal Fundus Image

Heart disease increases the mortality rate in the recent years across the world. So it is necessary to develop a model to predict the heart disease occurrence as early as possible with higher rate of accuracy. In this study, the cardiovascular disease is predicted by non-invasive method with the retinal image data. In this system, the retinal fundus image data are used to predict the heart disease occurrence. Cardiovascular disease can be detected from the changes in microvasculature, which is imaged from retina. The prediction of a disease is by considering features like age, gender, smoking status, systolic blood pressure, diastolic blood pressure, and HbA1c. Risk factors for heart disease occurrence are detected from the microvasculature of segmented retinal fundus image using MATLAB. The main objective of the proposed system is to predict the occurrence of heart disease from retinal fundus image with higher rate of accuracy.

R. Rekha, V. P. Brintha, P. Anushree

Chapter 72. Blind Speech Enhancement Using Adaptive Algorithms

Speech enhancement is used in all wireless telecommunication systems. Some of the interferences like white noise, periodic noises like human noise, and room reverberations may affect the speech quality. The delinquent of speech enhancement is communicated in this chapter. Least mean square (LMS), normalized least mean square (NLMS), and dual fast normalized least mean square (DFNLMS) are discussed in this chapter. The objective criteria are calculated for segmental signal-to-noise ratio, segmental mean square error, signal-to-noise ratio, and comparison results of the adaptive algorithms are discussed.

P. Shanmuga Priya, S. Selva Nidhyananthan

Chapter 73. Android Malware Detection

Smartphones and mobile tablets are rapidly becoming essential in daily life. Android has been the most popular mobile operating system since 2012. However, owing to the open nature of Android, countless malwares are intermingled with a large number of benign apps in Android markets that seriously threaten Android security. The botnet is an example of using good technologies for bad intentions. A botnet is a collection of Internet-connected devices, each of which is running one or more bots. The Bot devices include PCs, Internet of Things, mobile devices, etc. Botnets can be used to perform Distributed Denial of Service (DDoS attack), steal data, send spam and allow the attacker access to the device and its connection. To ensure the security of mobile devices, malwares have to be resolved. Malware analysis can be carried out using techniques like static, dynamic, behavioural, hybrid and code analysis. In this chapter, several machine learning techniques and classifiers are used to categorize mobile botnet detection.

Shymala Gowri Selvaganapathy, G. Sudha Sadasivam, Hema Priya N, Rajeshwari N, Dharani M, K. Karthik

Chapter 74. Test Data Compression Methods: A Review

Integrated circuit (IC) applications have become viable, reliable, and cheaper with deep-submicron technologies in VLSI industry. SoC (system on chip) is a microchip, which holds the necessary hardware and software to implement various functions onto a single chip. An IC should be tested during design process to check the correctness of the design and also to check the functionality of the design after fabrication. Testing cost shares almost half of the manufacturing expenditure. Achieving high-test quality in reduced VLSI geometry increases the complexity of the testing methods because of huge volume of test data. Automatic testing equipment (ATE) reduces testing efforts, but it has limited memory in comparison with the huge volume of the test data. One of the methods to reduce the burden of ATE is test data compression. The advantages of test data compression are as follows: (i) reduction in memory requirement for ATE and (ii) reduction in testing time. Though many advanced algorithms are used for testing of VLSI circuits, most of them are expensive in terms of test data volume and power. This survey discusses all the test data compression methods, and it helps the researchers in testing domain.

S. Rooban, R. Manimegalai

Chapter 75. Piripori: Morphological Analyser for Tamil

Morphology is the study of internal structure of word forms. A morphological analyser is a tool which gets the inflected, derived or compound forms of words as input and retrieves various grammatical implications of the input such as root word(s) and supporting morphemes. Prominent south Indian languages like Tamil, Telugu, Malayalam, Kannada and Tulu are agglutinative. In agglutinative languages, complex words are formed by joining morphemes together without changing the spelling or phonetics. With enormous amount of text data being generated daily in different platforms, it is important to have linguistic tools to analyse the same. Several approaches have been proposed for developing a Tamil morphological analyser. The approach adopted for Piripori Tamil morphological analyser is based on word-level morphological rules. This paper proposes a unique architecture for morphological analysis. The computational time taken by Piripori and the state-of-the-art morphological analyser have been analysed and compared with an input set of 137,144 words. It was observed that the proposed approach achieves faster computation when compared to the state-of the-art analyser.

M. Suriyah, Aarthy Anandan, Anitha Narasimhan, Madhan Karky

Chapter 76. A Comprehensive Survey on Strategies in Multicore Architectures, Design Considerations and Challenges

CMOS technology in contemporary period is enhanced with advanced features and compatible storage system. Advanced CMOS technology provides functional density, increased performance, reduced power, etc. System-on-chip (SoC) technology provides a path for continual improvement in performance, power, cost, and size at the system level in contrast with the conventional CMOS scaling. When a single processor is transformed into multicore processor, it faces a lot of hazards to confine the circuits into single chip. To emphasize the importance of multicore architecture, this paper provides a comprehensive survey on multicore architectures designs, constraints, and practical issues.

R. Radhika, N. Anusha, R. Manimegalai

Chapter 77. Blockchain-based e-Voting as a Service

Blockchain is one of the most innovative things in recent times in computing. It is a verifiable and immutable ledger of records. The rise of cryptocurrencies has brought them more focus. It is being made clear that the applications of blockchain are not only restricted to cryptocurrencies but also can extend up to the way we use our smartphones, the structure of security systems, banking systems, health monitoring systems, even e-governance, etc. In this paper, an attempt is made to show the importance of blockchain in e-governance, through e-voting. The impact of blockchain along with its synergy with e-voting has been discussed in this paper, which can be used for further investigations.

R. S. Shyam Prakash, G. R. Karpagam

Chapter 78. Optimum Resource Allocation Techniques for Enhancing Quality of Service Parameters in Cloud Environment

Cloud computing offers multitenancy with countless services and follows pay-and-use strategy. In this paper Quality of Service (QoS) parameters such as energy consumption and response time are considered for resource allocation. Customer satisfaction can be fulfilled by improving the QoS. Multi-Agent-based Dynamic Resource Allocation (MADRA) strategy, a multistage framework using QoS-based Resource Allocation (QRA) algorithm, and Artificial Immune System-Directed Acyclic Graph (AIS-DAG) model are proposed for optimum resource allocation. The performance of proposed approaches using the CloudSim toolkit is analyzed and compared. The experimental results show that proposed approach has high potential for the improvement in QoS and scalability by concentrating on resource and request validation.

M. Kandan, R. Manimegalai

Chapter 79. MIPGIOT: Monitoring and Improving the Productivity in Garment Unit Using IOT

Enhanced communication and increase in radio frequency improve the growth of emerging Internet of Things (IoT) applications. Textile manufacturers are always trying to improve the production and quality of the garments to sustain in the enormous competitive market. They perform enormous number of operations at different spots by the operators to determine the sustainability and profitability. All these activities need to be performed in synchronized and timely manner to achieve desired productivity. Smart/interactive textiles are one of the methods of deploying smart materials in textile sectors. Smart materials appear to think and memory to revert back to their original state and also communicate with the master system. Machines are deployed with various smart devices and they are made to interact with the main master system. These data are collected in a storage device, and they are transferred to master system for monitoring the user efficiently, by determining the machine’s ideal state. The master system deploys various data analytic algorithm for measuring the performance of the system when it receives the status of the machine. Depending upon the prior values, a model can be developed so that it can be used to reveal estimate time to complete the given task.

V. G. Prabhu, R. Manimegalai

Chapter 80. Text and Audio Transfer Using LI-FI Technology

Wired or wireless network has become the essential way to use the Internet and complete people’s task these days. As the number of people accessing the wireless network increases, it gets complicated to rely on an incorrupt signal due to the jammed airwaves. Though Wi-Fi delivers a speed up to 1300 Megabits per second, it is not adequate to hold all the users. To redress this drawback of Wi-Fi, a new concept is being introduced called as Li-Fi, which provides a speed up to 224 Gigabits per second. This is very analogous to Wi-Fi. The major disparity is that visible light is used by Li-Fi and radio frequency is used by Wi-Fi to transmit data. In this project an Android application is created to transfer binary data from one device to another device through Li-Fi, whereas for audio transfer, a hardware setup is implemented using LED, solar cell, and speaker.

S. Sabareeswaran, G. Madumitha, M. Shruthi

Chapter 81. Hyperspectral Image Segmentation Using Evolutionary Multifactorial Spectral Analysis for OMEGA Dataset

Mars region is being imaged with an exceptional combination of spectral and spatial resolution spectrometer using OMEGA instrument. The hyperspectral images of Mars provide spectral range and chemical species with high resolution. This paper presents a novel unsupervised segmentation algorithm named as evolutionary component analysis for remotely sensed hyperspectral image data for material identification in the spatial and spectral information. Sparse multinomial logistic regression (SMLR) algorithm is initially employed to learn the posterior probability distributions from the spatial and spectral information of the images containing class imbalance information to infer the class distribution of the testing hyperspectral data. Evolutionary multifactorial spectral analysis (ESA) helps to characterize noise and extremely mixed pixels with less training set with high training quality and utility with respect to spectral signatures and its spectral changes with less interaction for classification and end-member detection. The proposed segmentation approach based on ESA is investigated and estimated using both real and simulated hyperspectral datasets. ESA is evaluated for the endmember extraction in the mixed pixel revealing up-to-date performance when compared with advanced hyperspectral image classification techniques. The combined spatial–contextual information (ESA + SMLR) characterizes a state-of-the-art contribution in the research field of material identification. The proposed approach is exposed to present proper classification of the minerals of Mars surface in both the spatial and the spectral domain in short span of time.

Nagarajan Munusamy, Rashmi P. Karchi

Chapter 82. Novel Lifting Filter Bank for Bird Call Analysis

Lifting scheme is widely used in signal processing applications due to advantages such as fast implementation, less computational complexity, and perfect reconstruction (Cohen A, Daubechies, I, Feauveau J, Commun Pure Appl Math, 1992). The proposed work discusses the design of a Novel Lifting-based Filter Bank (NLFB) for analyzing bird calls and songs. The proposed filter bank has four lifting steps such as split, predict, update, and merge that are used for perfect decomposition and reconstruction. The proposed NLFB consumes 65% less area and 17% less power when compared to interpolated filter bank.

N. Subbulakshmi, R. Manimegalai

Chapter 83. Automatic Classification of Solid Waste Using Deep Learning

Solid waste management is an essential task to be carried out in day-to-day life. So an automated recognition system using deep learning algorithm has been implemented to classify wastes as biodegradable and non-biodegradable. Efficient segregation of solid wastes helps to reduce the amount of waste buried in the ground, thereby improving the recycling rate, and safeguards the soil from pollution.

V. P. Brintha, R. Rekha, J. Nandhini, N. Sreekaarthick, B. Ishwaryaa, R. Rahul

Chapter 84. Relevancy and Similarity Aware Drug Comment Classification Framework on Social Media Drug Posts

Social media offer superior platform for its clients to share their comments and opinions about their knowledge toward specific product. This attracts the medical investigators to acquire knowledge about drug products owing to user perspective. Moreover, gaining and analyzing knowledge about drug tweet status is extremely a complex talk. This is carried out in our previous investigation by introducing Dynamic Drug Data examination by Hybrid Transductive Support Vector Machine with Fuzzy C Means (DDDA-HTSVM-FCM) procedure whose primary objective is to carry out drug tweet classification to acquire users’ opinion. But the research lacks in terms of accuracy of opinion identification. Accuracy of classification might get diminished in prior work with the occurrence of more extraneous tweets aggregated in online. The classification recital might be lesser in the prevailing work with the arrival of huge volume of data with either labeled or unlabeled data. These crises are resolved in the anticipated research technique by introducing Relevancy and Similarity Aware Drug Comment Classification Framework (RSDCCF). In this research work, similarity-based filtering technique is provided to eliminate irrelevant reviews from a number of tweets accumulated online. It is performed by evaluating semantic similarity among verbal of sentences and score is allocated based on similarity level. Verbal with fewer score will be eradicated to raise learning performance and to diminish computation overhead which caused by processing irrelevant data. Then, Classification performance is enhanced with presence of mixed data by initiating ensemble classifier. Classifiers used in this work are Adaboost, SVM, Random Forest, and TSVM. The complete implementation of research technique is carried out in MATLAB simulation, and it is proven that anticipated technique offers optimal outcome associated with accurate prediction of user opinion in drug comments than prevailing work.

D. Krithika Renuka, B. Rosiline Jeetha

Chapter 85. Enhanced Particle Swarm Optimization with Genetic Algorithm and Modified Artificial Neural Network for Efficient Feature Selection in Big Data Stream Mining

High dimensionality would be one of the major challenges faced by people working in research with big data as a high dimensionality that happens, while a dataset comprises of a big number of features. For resolving this issue, often researchers make use of a feature selection step for identification and removal of irrelevant features and repetitive features. Acceleration Artificial Bee Colony-Artificial Neural Network (AABC-ANN) has been introduced in the preceding research for handling the feature selection process over the big data. Computational complexity and inaccuracy of dataset remain as a problem for these methods. Enhanced Particle Swarm Optimization with Genetic Algorithm – Modified Artificial Neural Network (EPSOGA–MANN) is described in the proposed methodology for avoiding the above-mentioned issues. Modules including preprocessing, feature selection, and classification have been included in this research process. Fuzzy C Means (FCM) denotes the clustering algorithm which is used to handle the noise information efficiently in preprocessing. Feature selection process is carried out by means of EPSOGA algorithm optimally in this research. More important and relevant features are selected by EPSOGA optimization algorithm and as a result more accurate classification results are achieved in this work for huge volume of dataset. Input, hidden, and output layers are the three layers of MANN. It is introduced for improving the time complexity by means of neurons. The performance evaluation of the research method is conducted in the Matlab simulation environment.

S. Meera, B. Rosiline Jeetha

Chapter 86. Feature Selection Techniques for Email Spam Classification: A Survey

In this digital world, most of the communication is done only through the Internet. Email is widely used for exchanging information not only for personal communication but also has an important part in business communication because of its effectiveness, fastness, and cost-effective mode of communication. Spam email is the serious problem on the Internet; when users click on to the spam mail, it starts spreading viruses in the user system, consumes lot of network bandwidth and email storage space, and steals user’s confidential data. Feature selection approach selects the best features from the dataset which removes irrelevant, redundant, and noisy data. The proposed paper offers email spam detection which incorporates various feature selection approaches like Information Gain, Correlation-Based Feature Selection, Genetic Algorithm, Ant Colony Optimization, Artificial Bee Colony, Particle Swarm Optimization, Cuckoo Search Algorithm, Harmony Search Algorithm, etc.; when classification is done after feature selection, it will enhance the performance of spam filtering.

V. Sri Vinitha, D. Karthika Renuka

Chapter 87. A Novel Paradigm Towards Exploration of Rechargeable WSN Through Deep Learning Architecture for Prolonging Network Lifetime

Rapid development of energy efficient technique for wireless sensor network and its proliferation can relieve the energy constraints on sensor to a little extent but limited the lifetime of the batteries in the sensor. Upon exploration, rechargeable wireless sensor network has potential to mitigate this issue by prolonging the network through extraction of renewable energy to replenish the sensor on the deployed region. Energy-constrained deep learning architectures have to be utilized as optimization objective to the existing routing protocol to enhance the reliability and flexibility of the network. In this chapter, a novel deep learning algorithm, named as deep belief network, has been proposed to achieve energy efficiency. The dynamic source routing protocol is been employed on this paradigm; the mobile sink is utilized for data gathering and replenishing of energy in the cooperative manner towards data transmission. The deep belief network exploits the route with shortest path through information of the mobile sink on a random transmission. Priorization is employed to sensor nodes with least energy will be charged by the mobile node while computation of the node density. Furthermore, the proposed model eliminates the localization issues and latency issue of mobile node. Moving trajectory of the mobile sink is determined with optimal velocity control mechanism. The simulation results of the proposed architecture proves that the proposed paradigm exhibits a good performance in terms of throughput, latency, packet delivery ratio, routing overhead, and energy utilization of nodes compared with state of approaches.

M. Ezhilarasi, V. Krishnaveni

Chapter 88. Why Feature Selection in Data Mining Is Prominent? A Survey

Feature selection is employed to diminish the number of features in various applications where data has more than hundreds of attributes. Essential or relevant attribute recognition has converted a vital job to utilize data mining algorithms efficiently in today’s world situations. Current feature selection techniques primarily concentrate on obtaining relevant attributes. This paper presents the notions of feature relevance, redundancy, evaluation criteria, and literature survey on the feature selection approaches in the different areas by many researchers. This paper supports to choose feature selection techniques without identifying the knowledge of every algorithm.

M. Durairaj, T. S. Poornappriya
Weitere Informationen