Information and Communication Technology for Intelligent Systems (ICTIS 2017) - Volume 2

This paper proposed the design and implementation of high speed communication link between two FPGAs using LVDS driver using 100 MHz clock. Initially Asynchronous and Synchronous communication are discussed and then synchronous communication is used for LVDS data communication. The speed of the communication is sensitive to the noise and sampling of the data at the receiver end. Therefore, the differential signal voltage level and sampling at the falling edge of clock are proposed in the design to maintain the high speed.

The need of recognizing individual from the low resolution non-frontal picture is hard hassle in video surveillance. In an effort to alleviate the hassle of popularity in low decision photograph, literature presents unique strategies for face recognition after converting the low decision photograph to excessive resolution. For this reason, this paper provides a method for multi-view face video notable decision using the tangential and exponential kernel weighted regression model. In this paper, a brand new hybrid kernel is proposed to carry out non-parametric kernel regression version for estimation of neighbor pixel within the first-rate decision after the face detection is done the usage of Viola-Jones algorithms. The experimentation is finished with the U.S. Face video databases and the quantitative results are analyzed the usage of the SDME with the prevailing strategies. From the result final results, we prove that the most SDME of 77.3 db is obtained for the proposed approach compared with the existing techniques like, nearest interpolation, bicubic interpolation and bilinear interpolation.

The internet in its broader sense has covered all dimensions of computer science and still continuing to grow dynamically with more subtle research possibilities and naïve implementation approaches. One such terminology in terms of balancing the load over a network of connected nodes is usage of mobile agents. Distributed applications, mobile devices and intermittent connections have fostered and provoked the need of the technology that support move of code and not of data. In this paper, the naive technique of load balancing that involves mobile agent paradigm is used. It promises to make full use of resources available within the network nodes based on their load handling capabilities.

In a large scale Wireless Sensor Networks (WSNs), designing an energy balanced clustering protocol has become a challenging research issues. This is due to fact that design of an energy-balanced clustering for maximizing the network lifetime of WSNs is a NP-hard problem. For solving this NP-hard problem, many meta-heuristic approach based clustering protocols are proposed in the recent years. However, these existing clustering protocols suffer from unbalanced energy consumption problem. In this problem, cluster heads are not uniformly distributed and overloaded cluster heads die out faster than under-loaded cluster heads. In order to solve this problem, we have proposed an energy balanced clustering protocol using particle swarm optimization called EBC-PSO. In the proposed protocol, we have used a novel multi-objective fitness function which contains three constraints such as average intra-cluster distance, residual energy and average cluster size. A detailed evaluation and performance comparison of the EBC-PSO with the three most popular protocols such as LEACH, PSO-ECHS, and E-OEERP are included.

Device to Device Communication (D2D) refers to a communication where devices communicate with each other directly without the need of cellular infrastructure. D2D is a new concept introduced under 3GPP; several researchers are working on providing solutions for D2D communication for the future network evolution such as IoT (Internet of Things) and IoE (Internet of Everything). Challenges associated with the D2D communication are security, interference management and resource allocation. Besides technical challenges, there are practical problems of incentivizing users to lend their devices to serve as relays for the traffic of others. To overcome these problems we have proposed a routing protocol for data packets of device communication called tree-based intelligent routing protocol based on SOFNET platform which results in support of IoT and IoE for the heterogeneous network. In this paper, we propose the architecture for routing in D2D communication. Simulation results on the proposed algorithm based on throughput are presented.

There are various destructive as well as non-destructive techniques available to detect corrosion in metallic surfaces. Digital Image Processing is widely being used for the corrosion detection in metallic surface. This non-destructive approach provides cost effective, fast and reasonably accurate results. Several algorithms have been developed by different researchers and research groups for detecting corrosion using digital image processing techniques. Several algorithms related to color, texture, noise, clustering, segmentation, image enhancement, wavelet transformation etc. have been used in different combinations for corrosion detection and analysis. This paper reviews the different image processing techniques and the algorithms developed and used by researchers in various industrial applications.

Word Sense Disambiguation has been a major challenge for various linguistic researches. Enough research has been carried in the past four decades. In machine translation, WSD plays a vital role in improving the accuracy of the translation. The automated translation of question papers from English to Hindi is one such key area which requires suitable WSD techniques to resolve ambiguity in a question word. When machine translates question sentences, it faces ambiguity problem that results in ambiguous translation. Identification of question type is important for remove ambiguity in the question paper. In this paper besides discussing WSD its approaches, resources for translation. We have also discussed question classification word sense disambiguation.

This paper proposes a hardware design, implemented on an FPGA, for a hybrid selective encryption and selective error correction coding scheme. FPGA’s are used as implementation platforms in image processing, as its structure exploits the temporal and spatial parallelism. The algorithm aims at implementing security and reliability in which encryption and encoding are performed in a single step using Bezier curve and Galois field GF (2m). The system aims at speeding up the encryption and encoding operations without compromising either on security or on error correcting capability by using selective encryption and selective encoding. The coding for hybrid crypto-coding algorithm is carried out using VHDL. The algorithm is simulated and synthesized using Xilinx ISE 10.1 software. The algorithm is implemented on Spartan 3 FPGA device 3s1000fg676-5. The proposed scheme reduces the hardware as modular arithmetic operations are involved.

PID controllers can be designed using analog and digital methods; digital PID controllers are most significantly used in the industries. FPGA based PID Controllers are preferred because of their improved settling time and are small in size, consume power efficiently and provide high speed of operation compared to software based PID controllers or microprocessor/microcontroller based PID controllers. The effort has been taken to implement the digital PID controller using FPGA device based on multiplier principle, which is implemented using MATLAB Simulink and system generator. The controller is designed for speed control of dc motor and implementation is accomplished on Xilinx Spartan 3 FPGA chip to achieve the settling time of 10 s this shows 5 s early settlement compared to basic PID controller. The resources consumption of the scheme is also presented.

Negotiation is an unconsciously has become a part of our life and we don’t even realize in our life that when we start negotiating. It starts from our childhood to adolescents from a chocolate to bike for getting good marks in the examination. But besides this we never realize that sometimes this acumen can help us becoming one of the successful negotiators in our professional life. Earlier in case of disputes the corporate houses prefer to have arbitration clause to avoid court hassle but now the companies are moving step forward to seek out the differences through negotiation that they are not turned into disputes with time. The communication technology has also helped the same by providing an e-platform in the form of electronically mediated negotiation. The paper is an attempt to discuss the relevance and steps of negotiation in commercial disputes and relevance of e-negotiation in the same lines.

Mobile ad hoc network (MANET) is a collection of wireless mobile devices that can communicate with each other and forming a temporary network without requiring any fixed infrastructure or centralized administration. The nodes of MANETs are always susceptible to compromise. In such scenario, designing a secure routing protocol has been a major challenge for last many years. In this paper, we propose Node Trust Path Trust Secure AODV routing protocol (NTPTSAODV) to enhance the security for MANETs through a new approach 2-tier trust based model for Intrusion Detections (IDs) techniques. To implement NTPTSAODV, Ad hoc On-demand Distance Vector (AODV) routing protocol has been modified for making it secure and thwart black hole attacks. In order to make result more accurate, performance of NTPTSAODV, AODV and black hole detection AODV (BDAODV) was tested in presence of multiple attacker cases and after observations of performance analysis, it can be concluded that NTPTSAODV is capable of delivering packets even in the presence of malicious nodes in the network. To evaluate the network performance, packet delivery ratio (PDR), Average End-to-End Delay (AED), Average Throughput (AT) and routing overhead are considered criteria.

India, a country where unity in diversity is practiced through several cultural, social, linguistic and religious adaptations in every facet of life. Here, communication between different states requires common platform or language of interpretation. Language happens to be a barrier in handling many societal issues including security of a state. The inter-state border activities reduce the efficiency of security deployments as the common goal of handling the miscreant is not available after he crosses the state border. So there is a need of natural language interface which can support different Indian languages. Any application catering to inter-state domain has a backend to process the information. Such databases use regional language to store and retrieve information. These regional languages are not user friendly for non-native users. Our goal is to design Natural Language Database Interface for the conversion of one of the Indian Languages i.e. Gujarati to English.

Digital media becomes an effective way of communication which is available round the clock to everyone including humans and machines. This put the requirement for machines to differentiate between human and machine as far as access of the website or its relevant services is concerned. CAPTCHA (Completely Automated Public Turing test to tell Computer and Human Apart) is a test that helps machines (or programs) to differentiate between human and machine. CAPTCHA should be easy for users to solve and difficult for bots to attack. In this paper, a simple and efficient text-based CAPTCHA verification scheme is proposed which is easy for human and hard for bots. The proposed scheme uses virtual keyboard, eliminates input-box, and does verification on the basis of the positions of the characters.

Collaborative Filtering (CF) is one of the most effective approaches to engineer recommendation systems. It recommends those items to user which other users with related preferences and tastes liked in the past. User-based and Item-based Collaborative Filtering (IbCF) are two flavours of collaborative filtering. Both of these methods are used to estimate target user’s rating for the target item. In this paper, these methods are implemented and their performance is evaluated on the large dataset. The major attention of this paper is on exploring different ways in which predictions from UbCF and IbCF can be combined to minimize overall prediction error. Predictions from UbCF and IbCF are combined through simple and weighted averaging and performance of these fusion approaches is compared with the performance of UbCf & IbCF when implemented individually. Results are encouraging and demonstrate usefulness of fusion approaches.

Digitization of data has now come in a big way in almost every possible aspects of modern life. Agriculture as a domain is no exception. But digitization alone does not suffice, efficient retrievability of the information has to be ensured for providing web services including question-answering. However, building an ontology for a vast domain as a whole is not straightforward. We view creation of an ontology as an incremental process, where small-scale ontologies for different sub-domains are expected to be developed independently, to be merged into a single ontology for the domain. The paper aims at designing a framework for ontology merging. The method is described with agriculture as the primary domain with several subdomains such as crop, fertilizer, as subdomains among others. The supremacy of the scheme over Protégé, a well-known ontology management software is demonstrated.

Financial time series clustering finds application in forecasting, noise reduction and enhanced index tracking. The central theme in all the available clustering algorithms is the dissimilarity measure employed by the algorithm. The dissimilarity measures, applicable in financial domain, as used or suggested in past researches, are correlation based dissimilarity measure, temporal correlation based dissimilarity measure and dynamic time wrapping (DTW) based dissimilarity measure. One shortcoming of these dissimilarity measures is that they do not take into account the lead or lag existing between the returns of different stocks which changes with time. Mostly, such stocks with high value of correlation at some lead or lag belong to the same cluster (or sector). The present paper, proposes two new dissimilarity measures which show superior clustering results as compared to past measures when compared over 3 data sets comprising of 526 companies. abstract environment.

Due to miniaturization of hardware and availability of low-cost, low-power sensors, Wireless Sensor Network and Multimedia Sensor Network applications are increasing day by day. Each application has a specific quality of service and experience requirements. The design of routing and MAC protocol which can fulfill the requirements of the application is challenging given the constrained nature of these devices. Considerable efforts are directed towards the design of energy efficient QoS-aware routing protocols. In this article, we present state of the art review of routing protocols for Wireless Multimedia Sensor Networks while addressing the challenges and providing insight into research issues.

The work is focused on detection of Obstructive Sleep apnea (OSA), a condition of cessation of breathing during night sleep caused by blockage of upper respiratory tract in an individual. ElectroCardioGram (ECG) signal is one of the clinically established procedures that can be relied on for deciding on the presence or absence of sleep apnea along with its severity in the subject at an earlier stage, so that the expert can advise for the relevant treatment. Earlier detection of OSA, can avoid the severe consequences leading to hypertension, Atrial-Fibrillation and day-time sleepiness that can affect the patient. ECG signal recordings from Apnea database from Physiobank, MIT website have been used for the purpose. The ECG signal based methods like QRS complex detection, RR interval variability, Respiratory Variability, Heart rate variability parameters used to detect OSA are compared and evaluated in order to select the most accurate method. Here we present the stepwise procedures, results and analysis of implementation methods used for detection of sleep apnea based on ECG signal using robust dataflow programming feature available in LabVIEW2014. Results indicate that accuracy, specificity and sensitivity of Heart Rate based detection method of OSA are 83%, 75% and 88% respectively and thus rated as one of the simple and reliable ways of detecting OSA.

Data cleaning perform in the Data Preprocessing and Mining. The clean data work of web server logs irrelevant items and useless data can not completely removed and Overlapped data causes difficulty during retrieving data from datasource. Previous paper had given 30% performance of datasource. So We have Implemented Smart Two-level clustering method to get pattern data for mining. This paper presents WebLogCleaner can filter out much irrelevant, inconsistent data based on the common of their URLs and it is going to improving 8% of the data quality, performance, Accuracy and efficiency of any Datasource.

IEEE 802.11-power save mode (PSM) independent basic service set (IBSS) Save, the time is divided into intervals of the signals. At the beginning of each interval signal and power saving alarm periodically all open windows (vocals). The station will be in competition with the rest of the frame window frame sent voice data leakage range. Element depends frame transmission IEEE CSMA/CA as defined in 802.11 DCF. A chance of transmit voice frames type of collision energy IBSS success. This article gives an analysis model with a chance of success output transmission window fixed size element. The results of the simulation analysis of the accuracy of the analysis.

This paper tells us about the smart process control laboratory in which the concept of industrial internet of things (IIOT) is implemented on laboratory-scale trainer kit. A case study of level control trainer is discussed along with implementation and results. PID control algorithm is implemented in order to control level of water in the tank. Today, IIoT is an emerging technology that brought the control and automation on the platform of IoT, i.e., the control and monitoring of sensors and actuator is done from remote location. This example is relatively for home automation where the mobile devices comes into the picture. The device challenges and requirements of the systems are discussed. The software platform chosen is NI LabVIEW through which the application data dashboard is used in mobile devices to communicate with the process computer.

Tumor identification and classification from various medical images is a very challenging task. Various image processing and pattern identification techniques can be used for tumor identification and classification process. Deep learning is evolving technique under machine learning that provides the advantage for automatically extracting the features from the images. The computer aided diagnosis system proposed in this research work can assist the radiologists in cancer tumor identification based on various facts and studies done previously. The system can expedite the process of identification even in earlier stages by adding up the facility of a second opinion which makes the process simpler and faster. In this paper, we have proposed a framework of convolution neural network (CNN), that is a technique under Deep Learning. The research work implements the framework on AlexNet and ZFNet architectures and have trained the system for tumor detection in lung nodules and well as brain. The accuracy for classification is more than 97% for both the architectures and both the datasets of lung CT and brain MRI images.

In this paper, Decision Tree (DT) based IDS is proposed for NSL-KDD dataset. The proposed work uses Correlation Feature Selection (CFS) subset evaluation method for feature selection. Feature selection improves the prediction performance of DT based IDS. Performance is evaluated before feature selection and after feature selection for five class classification (normal and types of attack) and binary class classification (normal and attack). The obtained result is compared and analyzed with the other reported techniques. The analysis shows that the proposed DT based IDS provides high DR and accuracy. The overall result for binary class classification for the dataset is higher than five class classification.

Due to advancements in domain of Information processing, huge amount of data gets collected which varies according to different time intervals. Structural models and Time-series models are used for analysing time series data. Time series models are very efficient as compared to structural models because modelling and predictions can be easily done. This paper gives a brief insight into Auto-regressive Models (AR), Moving Average Models (MA), Autoregressive Moving Average model (ARMA) and Autoregressive Integrated Moving Average Model (ARIMA). This paper also helps to understand the characteristics of the data which will be used for Time-series modelling.

In this paper, we are implementing an Intelligent Alarm Clock using skin detection approach and YCbCr color space model. This novel application is designed by keeping in mind the busy schedule and fast paced life of people living in this era. Focusing on the data and reports provided by various health organization, total time which is vital for comfortable sleep and peace of mind in an average adult is six or more than six hours on daily basis [1]. In the current scenario of hectic schedule of working in society, spare time of six hours still sound pretty compromising but, in reality, though, it is leading towards chronicle sleep deficiency. Deriving the inspiration from various research references, this alarm clock works on sleeping pattern of a person using Digital Image Processing. This work can be of great help for people suffering from sleeping disorders and hypersomnia. It is implemented using MATLAB. The tool used includes webcam, Image Processing Toolkit, and Image Acquisition Toolbox available in MATLAB.

String matching algorithms are among one of the most widely used algorithms in computer science. Traditional string matching algorithms are not enough for processing recent growth of data. Increasing efficiency of underlaying string matching algorithm will greatly increase the efficiency of any application. In recent years, Graphics processing units are emerged as highly parallel processor. They out perform best of the central processing units in scientific computation power. By combining recent advancement in graphics processing units with string matching algorithms will allows to speed up process of string matching. In this paper we proposed modified parallel version of Rabin-Karp algorithm using graphics processing unit. Based on that, result of CPU as well as parallel GPU implementations are compared for evaluating effect of varying number of threads, cores, file size as well as pattern size.

The massive amount of data, if publicly available, can be stored and shared securely for analysis and advancement. Mining of association rule besides classification technique is skilled of discovering useful patterns from big datasets. This technique results in the if-then form of rules and these rules are simple for end users to understand and easy for prediction. But it is apparent that the gathering and analysis of such data causes a serious menace to confidentiality and freedom. Hence, it interprets a field of privacy-preservation of data mining, which deals with efficient conduction and application of data mining without scarifying the privacy of data. This paper puts effort on the construction of class association rules generated by associative classification and applying privacy-preserving techniques on these rules to prevent its disclosure to the uncertified population.

Grown internet usage by individual and industries have also increased the attack vector in cyberspace rapidly. Botnet is a digital weapon used by attackers to commit cybercrime in stealthiest way for all type of illegal online activity. Botnet is well articulated attack responsible for many malicious activities in large volume and mass effective against any targeted organization such as confidential data theft, financial loss, distribution of pirated products, e-business extortion and network or service disruption. Because of its global nature of infection and innovative covert techniques of malware development to evade detection, it is also known as advance persistent threat. An analysis of this APT revealed the advancement in sophistication of bot malware by encryption methods, concealed network connections and silent escape as an effective tool for profit-motivated e-crime. Reverse engineering is procedure to analyze malware to classify its type, hazard, impact on machine, information outflow and removal of signature technique. Botnet (APT) detection needs improvised process to identify the channel, architecture and encryption weakness. In bot examination; Programming style, network protocol and behavior analysis can mitigate the APT by creating signature, prototype of behavior based approach and elimination of C&C servers. Reverse engineering is excellent way for defense the modern botnets to immune valuable information by identifying the evidence behavior, log collection and digital forensics. The main aim of study is to determine the most adequate approach to recreate a botnet incident. Network security is prime concern to avoid state sponsored attacks like botnet so security of digital nation and e-governance can be assured.

The ubiquitous growth of social networks opens a new line of research for developing algorithms and models for influence mining. Determining influential people in the network which consists of both positive and negative links between users is a challenging task. It becomes critical for businesses with fixed budget constraints to identify a group of influential people whose views will influence others’ behaviors the most. In this paper, we propose a model that aims to discover a set of influencers in signed social networks with maximal Collective Influential Power (CIP). We first construct an “influence network” between users and compute the influence strength between each pair of users by utilizing both the explicit trust-distrust information provided by users and the information derived from interactions between them. We then employ an elitist genetic algorithm that discovers a set of influencers with high influence spread as well as maximal enhanced joint influential power over the other users in the network. Experiments are performed on Epinions, a real-world dataset, and the results obtained are quite promising and clearly demonstrate the effectiveness of our proposed model.

Face recognition over aging is still a difficult but interesting problem in pattern recognition nowadays. It has many real world applications. It is highly affected with many uncontrolled parameters like variations in head pose, expressions and illumination. Aging also varies person to person, thus makes the task more difficult. This paper includes an approach proposed by us for solving this problem. Here, we introduced a novel feature descriptor that is a combination of Gabor and LBP features called as GLBP. We used Principal Component analysis (PCA) for dimensionality reduction and k-NN as a classifier. Proposed approach is experimentally tested on popular aging datasets FGNET and MORPH. It is observed from the experimental results that our approach is better in Rank-1 recognition accuracy as a performance measure.

Automatic text summarization can be defined as a process of extracting and describing important information from given document using computer algorithms. A number of techniques have been proposed by researchers in the past for summarization of English text. Automatic summarization of Indian text has received a very little attention so far. In this paper, we propose an approach for summarizing Hindi text based on semantic graph of the document using Particle Swarm Optimization (PSO) algorithm. PSO is one of the most powerful bio-inspired algorithms used to obtain optimal solution. The subject-object-verb (SOV) triples are extracted from the document. These triples are used to construct semantic graph of the document. A classifier is trained using PSO algorithm which is then used to generate semantic sub-graph and to obtain document summary.

In this paper, we propose a wireless sensor network based portable pollution monitoring system for monitoring the carbon monoxide (CO) concentration levels on the real time basis. Carbon Monoxide which is a critical and primary pollutant in air significantly affects the health of the people. With the rapid industrialisation and the exponential growth of automotive vehicles had led to the deterioration of air quality in the urban areas. Our design consists of Testbed of five nodes with calibrated carbon monoxide sensors for measuring CO concentration levels. By using the multi-hop mesh network, the CO sensors are integrated onto the Waspmote to communicate between the various nodes for the information exchange. The derived concentration levels of carbon monoxide from the different sensors on the board are made available on the internet through the platform which consists of Light Weight Middleware and Net Interface deployed on the server. Designed prototype had been implemented and tested in collecting the emission levels of CO in the Hyderabad city which had shown the consistent results under various circumstances.

In the wireless cellular networks in order to deal with irregular and expanding demand, channel must be allocated in such a way that spectrum is used efficiently, capacity is maximized with a minimum level of interference; this problem is called Channel/frequency Allocation Problem. The swarm intelligence category of Heuristic technique, i.e. Particle Swarm Optimization and Ant Colony Optimization for Hybrid Channel Allocation is investigated to find the optimal solution to the minimum interference. The fitness function designed is based on Graph Theory in PSO. The designing of fitness function is the probabilistic model with Sequential packing and ordering technique is explored with ACO. The interference level is represented by edges indicating co-channel and co-site. The signal to interference ratio is measured for Kunz benchmarks and the computation time is obtained. The performance of applied PSO and ACO is compared with the literature reported with Genetic algorithm (GA).

In this paper, pixel level features of the character are used for Devanagari numeral Recognition. The pixel distribution features for each numeral can be calculated after preprocessing the document image and converting it to binary. Based on these features the numerals are classified into appropriate groups. Histogram feature matching method gives erroneous results for the numbers like one and nine as they are having nearly similar histogram. In the proposed approach pixel distribution features are extracted in four directions. The overall performance of classification can be improved if more number of features is compared. The proposed approach gives improved results as compared to simple histogram matching criteria.

Through the progress in each hardware and software system technologies, automatic data creation and storage have become quicker than ever. Such data is called as a data stream. Streaming information is present everywhere and it’s usually a difficult problem to visualize, collect and examine such huge volumes of information. Data stream mining has become a unique experimental area in information finding because of the large size and rapid speed of data in the data stream, due to this reason conventional classification methods are not effective. In today`s a substantial amount of analysis has been done on this issue whose main aim is to efficiently solve the difficulty of information stream mining with concept drift. Class imbalance is one of the problems of machine learning and data processing fields. Imbalance data sets reduce the performance as well as the overall accuracy of data mining methods. Decision making towards the majority class, which lead to misclassifying the minority class examples or moreover considered them as noise.

Rough Intuitionistic Fuzzy C-Means Algorithm is a combination of Fuzzy Sets, Rough Sets and Intuitionistic Fuzzy Sets. This algorithm provides high quality clustering over numeric datasets. However, RIFCM is highly inconsistent over Image datasets. In this paper, we combine RIFCM with Firefly Algorithm. Firefly algorithm is a meta-heuristic bio-inspired algorithm which mimics the behavior of fireflies. Our experimental results prove that using Firefly algorithm before RIFCM lends stability to the clustering output and considerably reduces the number of iterations required for convergence.

This paper discusses affix removal and statistical based Stemming algorithms in detail with stemmer-generated output from some Standard English text and dictionary. Comparative empirical studies of all these stemmers are also discussed here with respect to number of stem token generation from single root morphed word variants and computation time. First part of the paper deals with introductory discussion of stemming and lemmatization. Second part of the paper focuses on algorithms of affix and statistical based stemmers with their empirical output. Last part describes the steps of the comparative tool for the same. Finally conclusion section wraps up whole discussion about stemming. This paper can assist researchers working in the field of text mining.

Customer churn prediction has always been an important aspect of every business. Most of the companies have dedicated churn management teams which work for both churn prevention and churn avoidance. In both of the scenarios it is highly required to identify customers who may change their service providers. In this paper we have tried to propose a neural network based model to predict customer churn in telecommunication industry. We have than used Artificial Bee Colony (ABC) algorithm for neural network training and observed a substantial improvement in accuracy. To prove the efficacy of our model we have compared it against Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Ant Colony Optimization algorithm (ACO). Simulation result shows that ABC trained neural network is more accurate than others in predicting customer churn in telecommunication sector.

A novel approach for dynamic decision making in retail application by expertise based cooperative reinforcement learning methods (ECRLM) is proposed in this paper. Different cooperation schemes for cooperative reinforcement learning i.e. EGroup scheme, EDynamic scheme, EGoal-oriented scheme proposed here. Implementation outcome includes demonstration of recommended cooperation schemes that are competent enough to speed up the collection of agents that achieves excellent action policies. This approach is developed for a three retailer shops in the retail market. Retailers be able to help with each other and can obtain profit from cooperation knowledge through learning their own strategies that exactly stand for their aims and benefit. The retailers are the knowledge agents in the hypothesis and employ reinforcement learning to learn cooperatively in situation. Assuming significant hypothesis on the dealer’s stock policy, refill period, and arrival process of the consumers, the approach is modeled as Markov decision process model thus making it possible to apply learning algorithms.

Speaker Diarization deals with determining “who spoke when?” with the help of computers. It is extremely useful in speech transcription, subtitle generation and extracting opinions among others. Diarization is a process in which a relatively long audio recording is processed and speech segments are labeled using respective speaker identities. Such systems are also helpful in determining number of speakers in any conversation or meeting. This field of research is active since long and researchers have been successful to improve the system over time. In this paper, we have made an attempt to study the effect of varying window sizes and threshold criteria on performance of speaker diarization system. Experiments are conducted using LIUM (Laboratoire d’Informatique de l’Universite du Maine) toolkit on Augmented Multi-party Interaction (AMI) meeting data corpus. The proposed approach has shown promising results in case of significantly less number of frames per window.

Wireless Sensor Network is made of tiny energy-constrained nodes with a limited amount of communication, computation and storage capabilities. Also, it is inconvenient to change batteries of the sensor nodes due to large-scale deployment in hostile environments. Hence, network longevity becomes the prime concern for WSNs. This paper presents Distance and Energy based Advanced LEACH protocol named DEAL. DEAL considers energy and distance of a node during the cluster head election process. Simulation results show that DEAL enhances the stability period and slashes the instability period as compared to ALEACH protocol.

This paper presents a comparative analysis of Digital Cosine Transformation (DCT) technique and Discrete Wavelet Transformation (DWT) technique in digital image watermarking. We have used standard digital images for analysing of Watermarked Images and applied standard attacks on the watermarked image. An experimental comparison was made using Matlab.

In the escalating trend of atomization and online information, text summarization bolster in perceiving textual information in the form of summary. It’s highly tedious for human beings to manually summarize large documents of text. In this paper, a study on abstractive and extractive content rundown strategies has been displayed. In Extractive Text Summarization it talk about TF-IDF, Cluster based, Graph theory, Machine learning, Latent Semantic Analysis (LSA) and Fuzzy logic approaches. Abstractive rundown techniques are ordered into two classes i.e. Structured based approach and Semantic based approach. In Structure Based approach it talk about Tree based, Template based, Ontology based, Lead & Phase based and Rule based method. In Semantic Based Approach it talks about Multimodal semantic, Informative item based and Semantic graph based method. The central idea of this method has been elaborated further, apart from idea, the advantages and disadvantages of these methods have been procured.

Internet of Things (IoT) enable humans to get higher level of automate by developing system using sensors, interconnected devices and Internet. In ICU, patient monitoring is critical and most important activity, as small delay in decision related to patients’ treatment may cause permanent disability or even death. Most of ICU devices are equipped with various sensors to measure health parameters, but to monitor it all the time is still challenging job. We are proposing IOT based system, which can help to fast communication and identifying emergency and initiate communication with healthcare staff and also helps to initiate proactive and quick treatment. This health care system reduces possibility of human errors, delay in communication and helps doctor to spare more time in decision with accurate observations.

Now a days cloud computing is the major area of research Because cloud computing has own many benefits. Cloud computing also provides cost effective Resources so that it can become more and more helpful to IT trends. Distributed computing is an extensive arrangement that conveys IT as an administration. It is an Internet-based registering arrangement where shared assets are given like power disseminated on the electrical grid. Cloud suggest to a particular IT environment that is expected with the end goal of remotely provisioning versatile and measured IT resources. Whereas the Internet gives open access to many Web-based IT assets, a cloud is commonly exclusive and offers access to IT assets that is metered following SLAs implementation depends on guidelines that are redesigned in runtime so as to proactively recognize conceivable SLA Violations and handle them in a proper manner. Our proposed framework allows the creation and implementation of effective SLA for provisioning of service. SLA management are one kind of common comprehension between CSP (Cloud Service Provider) and customer.

To cater the need of a user as the requirement of infrastructure, application building environment or contemporary software private and public clouds are exiting for provisioning of such services. But cloud providers are facing the problem of how to deploy their applications over different clouds keeping in mind their different requirements in terms of QoS (Cost, resource utilization, execution time). Different clouds have different advantages such as one cloud will be more reliable and efficient whereas, private cloud will be more secure or less expensive. In order to get the benefits of both clouds, we can use the concept of cloud federation. When using cloud federation it becomes important how to schedule large workflows over federated clouds. Proposed work addresses the issue of scheduling large workflow over federated clouds. SMARTFED is used for the cloud federation and our algorithm is used to schedule different workflows according to the QoS parameters over the federation.

This paper proposes a method for change detection based on Modified log ratio. The ratio of variance and mean has been considered along with the logarithmic ratio of the pixels in the images. The ratio of variance and mean played an important role in balancing the preservation of details and robustness to noise in difference image. The multi-temporal images are compared through the proposed method and a difference images has been generated. Fuzzy c means clustering (FCM) has been used to classify the changed and unchanged areas. The results are compared based upon various parameters like false negative (F_n), false positive (F_p), Kappa coefficient (K_c) and percentage correct classification (PCC). The qualitative and quantitative results show that the proposed method offers less speckle noise, higher accuracy and Kappa coefficient value as compare to the other direct comparison based algorithms.

We have presented an idea in this paper for detecting and tracking topics from news articles. Topic detection and tracking are used in text mining process. From data which are unstructured in text mining we pluck out information which are previously unknown. The objective of this paper is to recognize tasks occurred in different news sources. We are going to use agglomerative clustering based on average linkage for detecting the topics, calculate the similarity of topics using cosine similarity and KNN classifier for tracking the topics.

The journey of Cellular Mobile Network could be considered from rags to riches. The base-band of Cellular network and its working lies in its proper planning, dimensioning and deployment. Some of the dimensioning techniques for the next generation mobile wireless network have been discussed in this paper. The research has been focused for (Next Generation) LTE-A (Long Term Evolution-Advanced) Network Planning, and dimensioning of the Heterogeneous Network. It is divided into three categories. Firstly, planning based on various radio network parameters (like Quality of Service, Cost, power and path-loss) under consideration. Second part comprises of network planning based on various issues and corresponding solution. Various Deployment strategies using Voronoi Tessellation, AHC (Agglomerative Hierarchical Clustering) and Traditional K-means Deployment (TKD) methods have been discussed. Third part shows the importance of planning for deployment of the network with demography. Macro network, urban scenario with moderate density has been considered for dimensioning.

Identification of emotions from speech is a system which recognizes the particular emotion automatically without basing on any particular text or a particular speaker. An essential step in emotion recognition from speech is to select significant features which carry large emotional information about the speech signal, speech signal has an important features. The features extracted from the shape of speech signal are used such as MFCC, spectral centroid, spectral skewness, spectral pitch chroma. These features have been modeled by Gaussian mixture model and optimal number of Gaussians is identified. IITKGP-Simulated Emotion Speech corpus is used as database and four basic emotions such as anger, fear, neutral and happy are considered. The different mixture of spectral features is extracted and experiments were conducted.

To withstand the quick development with the demand for energy and the overall demand cost, enhanced effectiveness, dependability and adaptability, smart strategies should be carried forth in energy sector for our earth and vitality protection. In the electrical domain DSM can be a part of smart grid where the consumers can participate themselves to decrease the peak load and eventually the load profile can be reshaped. A portion of the DSM method is peak clipping, load shifting, valley filling and energy conservation. The paper involves the concept of load shifting to low voltage consumers using several types of appliances and in large numbers. Load shifting with respect to day ahead forecast is formulated as a minimization problem and are solved using learning based evolutionary algorithm. Simulations were carried out with a specific test case using Mat Lab and the results show a substantial peak reduction and cost savings for the future smart grid.

Using Sentiment analysis one can understand interaction of a user with the movies through their feedback. Here analysis is done based on the movie reviews that can be collected from many sources. Twitter is one among the foremost frequent on-line social media and micro blogging services. Due to the popularity of twitter it has become a useful resource for collecting sentiments through API or other data mining techniques. Our work here presents an examination on the evaluation of the machine learning algorithms (Random Forest, bagging, SVM and Naïve Bayes) in R together the public opinion for example opinion about ‘Civil War’ Movie. Here we have used ‘Random Forest’ to show its better performance in the analysis of movie reviews.

Social networks have become a “household name” for internet users. Identifying shortest paths between nodes in such networks is intrinsically important in reaching out to users on such networks. In this paper we propose an efficient algorithm that can scale up to large social networks. The algorithm iteratively constructs higher levels of hierarchical networks by condensing the central nodes and their neighbors into super nodes until a smaller network is realized. Shortest paths are approximated by corresponding super nodes of the newly constructed hierarchical network. Experimental results show an appreciable improvement over existing algorithms.

Today there exist various sources that provide information in very massive amount to serve the demand over the internet, which creates huge collection of heterogeneous data. Thus existing data can be categorized as unstructured and structured data.In this paper we propose an idea of a tool which intelligently preprocesses the unstructured data by segmenting the whole document into number of sentences, using deep learning concepts with word2vec [11] and a Recurrent Neural Network [13]. At the beginning step we use word2vec which was introduced by Tomas Mikolov with his team at Google, to generate vectors of the inputted text content which will be further forwarded to Recurrent Neural Network. RNN takes this series of vectors as input and trained Data Cleaning Recurrent Neural Network model will perform preprocessing task (including cleaning of missing, grammatically incorrect, misspelled data) to produce structured results, which then passed into automatic summarization module to generate desired summary.

English to Hindi Translator using Translation Memory is a translation tool that transforms an English sentence to its proper Hindi meaning. This translator works on both exact match and fuzzy match. When the input source in English is divided into segments and is completely matched with the database then the appropriate translation for that input is directly fetched from the database. The matching process between the input and the source file is done with the help of Edit Distance Matching algorithm. The case when the input is not 100% matched with the database that is considered as the case of fuzzy match so; in that case N-gram modeling is performed in order to give an appropriate translation to the user. The case in which the input is not completely matched with database (fuzzy match) score is calculated to get the match percentage. The case when the input in the English language is not matched with the database at all, then an appropriate algorithm is used that gives word to word translation of that particular input.

This paper provides the method and process to build machine learning system using Deep Neural Network (DNN) for lexicon analysis of text. Parts of Speech (POS) tagging of word is important in Natural language processing either it is speech technology or machine translation. The recent advancement of Deep Neural Network would help us to achieve better result in POS tagging of words and phrases. Word2vec tool of Dl4j library is very popular to represent the words in continuous vector space and these vectors capture the syntactic and semantic meaning of corresponding words. If we have a database of sample words with their POS category, it is possible to assign POS tag to the words but it fails when the word is not present in database. Cosine similarity concept plays an important role to find the POS Tags of the words and phrases which are not previously trained or POS Tagged. With the help of Cosine similarity, system assign the appropriate POS tags to the words by finding their nearest similar words using the vectors which we have trained from Word2vec database. Deep neural network like RNN outperforms as compare to traditional state of the art as it deals with the issue of word sense disambiguation. Semi-supervised learning is used to train the network. This approach can be applicable for Indian languages as well as for foreign languages. In this paper, RNN is implemented to build a machine learning system for POS-tagging of the words in English language sentences.

The present study is an attempt for exploring parameters which represents the overall performance of Small Scale Wood Industries (SSWI) in Nanded district using statistical technique. This study specially focuses on some important parameters viz. availability of human resources, financial, production, transportation and marketing management aspect. To study the various in sights of the SSWI based on the focused parameters. The importance of this study on both the academic and the application levels is attributed to SSWIs, despite their contributions to the economy, have not been given due attention as the research of performance has been biased towards large enterprises.

This paper provides with the exposition of how robotics can offer to various places of agriculture. As we know that agriculture is the backbone of our country. We can make improvement in agriculture by replacing man’s power by using latest technologies like robots. The main motive of our prototype is to help farmers in various agricultural operations like spraying of pesticides and fertilizers by using the solar energy. As we know that pesticides and fertilizers contain harmful chemicals which affects human body. This prototype will provide safety to farmers, precision agriculture and high speed. The cost of our prototype is effective as it works on the solar energy (Renewable energy) and it is based on the wireless technology (RF communication) by using Embedded system.

In this paper we describe an Intelligent System for Automatic Transfer Grammar creation using parallel corpus of source and target language. As we know about English and Hindi language, the structure of Hindi is Subject-Object-Verb (SOV) while in English the structure is Subject-Verb-Object (SVO). Now the system has to decide in which order to translate the given source language (English) to the given target language (Hindi). The grammatically parsing source sentence has to generate target language on basis of grammar rule which we have created based on parallel corpus. These unique rules are always applicable when the same input structure is found and regenerate the output on basis of the grammatical rules. Thus it gives better accuracy in terms of quality of translation. Reordering [1, 2] is important part of Transfer Grammar which is very helpful in language pairs of distant origin. We focus on designing a system that gives correct reordering for English-Hindi Machine Translation System including simple, compound as well as complex sentences. Reorder sentences gets generated based on probability. The system is evaluated on the basis of precision, recall and f-measure.

Cloud computing is a facsimile for enabling ubiquitous, on-demand access to a shared pool of configurable reckoning resources which can be rapidly provisioned and released with minimal management effort. The services accorded by cloud are benevolent to many patrons. They are many new fangled ways to exert services of cloud but the prominent thing here is transcendence of service of service exerted by patron and the time spend by the patron for using the service. In cloud computing it is very important that patrons should be able to exert the service with exquisite transcendence and also patron should exert the service that he desires without waiting for long time. So the concept induced in this paper is new-fangled method for forthcoming patrons to exert service of desired transcendence and service which has fewer loads among the services available in cloud. Rough Clairvoyance Fuzzy C-means clustering (RCFCM) algorithm is exerted for clustering the services based on service transcendence by congregating feedback testimony from patrons who exerted the service. This RCFCM algorithm helps in giving testimony to forthcoming patrons regarding service transcendence of services available in cloud. While congregating feedback testimony and storing feedback testimony lot of security, fidelity issues arises, so in this paper decision trait is included therefore only valid feedback testimony from patrons is cogitated. Collateral method provides security while congregating feedback testimony. If patrons know which service transcendence is best then everyone tries to ingress only the services with exquisite transcendence and load of services with exquisite transcendence increases. As load increases again it takes lot of time for the patrons to access the service to solve this load predicament ad-judicature is exerted. Unfeigned and proficient co-conspirators are recruited for accomplishing complex and secure tasks in methodology by using nodular method. Therefore the methods induced in this paper are benevolent for the forthcoming patron’s to gain erudition about services transcendence and to exert desired service by patrons that having less load among available services in cloud so that patrons feel ecstatic and satiated by using the cloud services.

Horticulture includes tropical and subtropical fruits, vegetables, spices, flowers, medicinal and aromatic plants. Horticulture sector is a major growth of Indian Agriculture. India is second largest producer of fruits and vegetables in the world. But the post-harvest loss is because of weak supply chain entities like storage facilities, bad transportation facility, market facility, and not proper packaging, not use of modern techniques, not proper post-harvest management. Due to post harvest loss actual need of fruits does not satisfy and so that need to import the fruit from outside the country. If import of fruit is higher than the export then it will impact on balance of payment of India, value goes negative. Post-harvest loss indirectly affect on our Indian Economy. By using modern technologies post-harvest loss can be reduced. For example, Geographic Information System (GIS) can be used for analysis of spatial data and helps also in decision making in problem. Location based recommendation system will also help to recommend the location of cold storages and establishment of new cold storages.

With increasing boom in the tech market many online food delivery systems have come up but almost all of them have some or the other flaw such as restriction on orders or extra charges. Hence, the work presented in this paper proposes use of enhanced version of SET protocol and a delivery protocol which, aims at saving conventional resources such as human resources and fuel and also time by using pool delivery mechanism.

The many advantages of increase in Human Machine Interaction are obvious but it has also led to issues such as emotional imbalance, depression, reduction in interpersonal communication etc. Internet of Emotions can be broadly categorized as internet based technologies which aim to mitigate these problems and facilitate better Human to Human interaction in real world. IoE can be defined as an ecosystem where emotion packets travel via internet to manage user’s real time experience. We propose a system which will detect emotional state of the user, categorize it and actuate outer net elements to manage the emotion of the user. Detailed algorithm is given which includes use of the passive sensors, smartphone, big data analytics and machine learning. The framework is further explained with example of stress management. The proposed system based on affective computing will play a vital role in development of products and platforms which emphasises user involvement.

Specialty of Mobile Adhoc Networks (MANET) is that in current mobile technology it is the most promising and highly developed elucidation due to its remarkable performance in offering network connectivity even in very radical situations of adversity where there is maximum chance of link failure and more necessity of quick set up of networks. The elemental routing procedure in a MANET involves facilitating unremitting communication in the network system between two mobile stations during required period of time and the basic factor being selection of the most suitable forwarding node to proceed the real time packets from source towards destination in such a way that the optimization of the network can be achieved by maximum utilization of available resources. Considering that real time applications are one of the most challenging issue in MANET, due to transportation of high volume of data including audio, video, images, animation and graphics, this paper presents a monitoring approach for checking the Quality of Service (QoS) conditions during competent routing using the concept of Mobile Agents. An intelligent mobile agent is designed in the proposed QoS platform that monitors and controls the QoS processing tasks using longest critical path method at the forwarding node to select it as the best option out of all neighbor nodes. Simulation results shows higher packet delivery ratio and comparatively reduced bandwidth consumption overhead when it is compared with other similar approaches.

Personalized education in an ICT enabled environment is a contemporary matter today. The adaptation of education to diverse types of student is becoming a big challenge. Proposing personalized learning in the digital era, we obtained a new dimension called an agent based adaptive learning. This has led to adopt knowledge management practices that provide innovation in knowledge clustering for active learning. In the ICT enabled learning, the learners are geographically scattered. Agent-Oriented System simulates the teaching-learning pedagogy by sensing the environment, listing several traits, observing the user behavior, finding pattern and learning pace of the learner. This leads towards imparting intelligence to an agent that helps both learner and the tutor to build a smart teaching-learning environment. This paper proposes an agent which works as a middleware which uses semi supervised learning mechanism with both forward and backward chaining for the inference to impart intelligence in learning environment.

Security is one of the most important challenge that users face in migrating to cloud services. Users will lose the direct control over their data and they need to trust the cloud service provider (CSP) for security and access control. Therefore, this raises concern about security and privacy of data. Due to vulnerability of data being stored at these super data center, leads to more introspection about security in cloud computing. Thus, the security is the more prominent concern that need to be addressed to provide safe services. Many researches are going on for improving the security of the cloud storage systems. This survey paper focus on the trusted storage cloud computing proposed by different researchers.

Component Based Software Engineering (CBSE) is one of the specialized methodologies in the process of developing the software. The motivation behind the CBSE is software reuse that is using off the shelf components. The software component in reuse may be a design, document or a piece of code. The components considered in this paper are source codes, in particular functions. In order to have an efficient reuse of the components, they are to be described effectively and clustered so as to retrieve the components with a minimum effort. This paper shows the description of a software component based on the facets. An important facet in the description of software component is the behavior. The behavior is extracted from the comment lines present in the code, later these comments are converted to first order predicate.

Phishing is a fraudulent way that is used to entice innocent users to a scheming website by employing legitimate looking email and messages for illicit purposes. In this paper, a brainstorming study on vulnerability causing email phishing is determined. These vulnerabilities are spared in three categories on the basis of email structure i.e. Page-content vulnerability, Domain vulnerability and Code-scripting vulnerability. Here, an Anti-Phishing Effectiveness Evaluator Model (APEE Model) is proposed to examine the effectiveness of existing Anti-Phishing Mechanism. The implementation of this Anti-Phishing Effectiveness Evaluator Model (APEE Model) is done with on existing Anti-phishing mechanism. Their effectiveness is evaluated and a result metric is listed. In this research paper, Major finding are reported which could lead the recent researcher to deliver effective Anti-Phishing Solution.

The face is the fundamental basis for the identification/recognition of a person. Recognition of a individual by the face data and methods to extract the unique facial feature points from the facial image have been significantly increased during the last decade. The periocular region has become the powerful alternative for unconstrained biometrics with better robustness and high discrimination ability. In this paper, multiple descriptors are used for deriving the discriminative features of the periocular region and city block distance is used to compute the similarity between the feature vectors. Feature extraction techniques employed in recognition using the periocular region are Local Binary Patterns (LBP), Local Phase Quantization (LPQ) and Speeded Up Robust Feature (SURF). Results of Periocular modality are then compared with the results of face patterns. Periocular region showed significant accuracies compared to face with only using 25% of the full face. Experimentations are carried on FRGC database, and accuracies of both the periocular and face regions are compared.

Improving quality of student is related with increasing their learning and skills in the concerned field. The benchmark used to measure the learning and skill is obtained score of student and participation in various activities. During a session, this session may be a semester in short term and in longer term it may be program duration, student generates huge size of data. If this data is studied properly then learning and skill easily grouped. After knowing a person’s weakest zone a customized solution will provide to nurture him/her in required area of development. In this paper data will be studied to identify weaker student so that customized tutoring offered to him/her.

The latest internet has become a collaboration and communications platform, in that e-mail system is one of the most reliable internet services. Sending a spam e-mail is an economically useful commerce for intruders, with the very good earning of millions of dollars. The spam e-mail has become a critical issue to web and society, to stop/reduce the spam e-mails filtering techniques is not sufficient. This paper proposes to recognize spam domain by reading spam e-mails. These spam domains are nothing but Uniform Resource Locator (URL) of the website that intruder is promoting. The approach is based on extracting mail content; links from URL injected e-mail and subject of spam e-mails. These extracted parameters are grouped together through clustering algorithms and evaluated. This proposed work can be help as additional accessory to already available anti-spam tool to recognize intruders.

This paper shows an electronic route framework for outwardly weakened and daze individuals (subject). This framework comprehends impediments around the subject up to 500 cm in front, left and right course utilizing a system of ultrasonic sensors. It successfully figures separation of the identified protest from the subject and plans route way likewise keeping away from obstructions. It utilizes discourse input to mindful the subject about the recognized deterrent and its separation. This proposed framework utilizes ATmega16 microcontroller based implanted framework to process continuous information gathered utilizing ultrasonic sensor organize. In view of bearing and separation of identified hindrance, important pre-recorded discourse message put away in APR9600 streak memory is conjured. Such discourse messages are passed on to the subject utilizing headphone.

In this paper we have collected the data from selected population of Rajkot city by the way of android and iPhone application after collecting all radio signals data like Wi-Fi signal power, GPS signal power, 4g signal power, 3g signal power, and Signal to noise ratio in different mobile device in different geographical location we can apply datamining technique by which can measure the different type of the scenario. After applying different method we can find hidden pattern and many insight to deal with interference situation.