Information and Communication Technology for Intelligent Systems (ICTIS 2017) - Volume 1

In this paper monitoring and surveying of data along with its protection and security parameters through the continuous evaluation of different OSNs are performed sequentially to identify any focused persons availability and activity in multiple OSN channels. A real time scenario in OSNs domain is initially monitored for data extraction of publicly available data. The available data is matched using semantic matching algorithm to ensure the link establishments, data consolidation and the data segregation which could possibly reveal some details and also used for the purpose of study, research and strategic report making. The main theme is the forensic evidence gathering and also enlisting the security parameters of different OSNs.

Possibly the ultimate valuable utility of accurate personal detection is, providing security to authorized access systems from awful security attacks. Within all the currently present biometric schemes, fingerprint identification is most widely used from past few decades. Fingerprint identification scheme is very popular but there are some complexities in recognizing the patterns. Like: Greater dislocation or rotation in image which can causes overlap between fingerprint and create problem during matching, Distortion in fingerprint image, Skin condition and amount of pressure at the time of enrollment and matching, Feature extraction problems. This research paper proposes a cooperative biometric multimodal for fingerprint identification based on minutiae matching by addressing above mentioned problems. This research paper proposes a design a fingerprint identification technique which is expected to give better accuracy and acceptance rate. Main aim of this research work is to decrease False Acceptance Rate (FAR).

With the ever-increasing population of the world, increase the accidents at an alarming rate. Accidents are the most uninvited and unintentional miss-happening that causes a lot of injury and loss to human life. So there is an urgent need to stop these accidents to save mankind from destruction. Deaths by accidents have increased and have become the second largest reason for death in the world. Keeping in mind the status of accidents, a working hardware prototype has been proposed along with android based application to avoid road accident. A technique has been suggested where collision of vehicles can be avoided by alerting the driver with a buzzer and a message. Position of the approaching vehicles can be seen on an android application installed on a mobile device that aims to depict the exact location of the vehicles on Google map This whole proposed method is different from other methods as the methods that have been invented until date are mostly in-built techniques while the method proposed here is a prototype that can be implemented in real world.

Regression Models evolved since in Health Diagnosis contributes to predict the health condition through various approaches and analysis. These analytical approaches require statistical methodology for the better diagnosis in data mining applications, utilizing numerical predictions through regression coefficients’ approximations. Here, we propose a model to diagnose the Health condition through error pattern discovery from the known and unknown databases; by designing a hidden layer to extract hidden patterns through Hidden Decision Tree approach. Our proposed model evaluates the database for (a) resulting the different health diagnosis, (b) patternising these to optimized queries based on returned results, (c) effective implementation and extraction of data being supplied by the end users, to handle data regularities and irregularities and (d) evaluate the pattern through HDT approach for the accuracy of the knowledge discovery. Proposed model can be implemented on a medical data to evaluate the error pattern for numeric additive quantities to obtain a better health diagnosis model.

Tenderfoot, presently clients who are utilizing the web however do not worry about the security issues. The information that is being transmitted on the system is not thought to be protected. There is such a variety of dangers like sniffing, ridiculing, phishing exits. With the assistance of a few devices like Wireshark, firewall and Microsoft disk operating framework, we can counter quantify the assaults. Here, in this paper we proposed an answer, which is greatly, improved the other proposed solutions based on the ARPWATCH and ARP central server (ACS).

The process of adding significant information of two source images obtained from various sources into one image is called image fusion. Large volumes of data informations are obtained from various remote sensors. These informations are useful for image diagnosis through image fusion. Thus image fusion is the promising area of research. Many methods of image fusion have been suggested by the previous authors to produce a fused image having higher spatial resolution, but due to large amounts of data calculations, it is a time-consuming process. Therefore, a reconfigurable hardware system having high speed such as Field-programmable Gate Array (FPGA) is used for solving complex algorithm with reduced computation time to achieve parallel operation with high-speed characteristics. This paper describe the design and implementation of improved speed discrete wavelet transform based multisensor image fusion process with its implementation on hardware. MATLAB 2016a Simulink tools are used to integrate the Xilinx System generator with averaging method for image fusion. Algorithm design has been synthesized in Xilinx ISE 14.1and the same is implemented on ML 605 Virtex-6 FPGA kit. From the result, it is observed that the design consumes a total power of 4.36 W and operates at a maximum frequency of 851. 06 MHz.

With the proliferation of so many Bschools in India, it is becoming more and more difficult to choose a good Bschool. Over the decades since management education rankings first appeared, numerous debates have surfaced about their methodologies and objectivity. Although there has been significant research, especially by the various coaching institution magazines, about the ways in which rankings might be improved, there has been less research on providing the rankings according to the preferences of a person. In addition, there has been scant research on how rankings may impact students’ access to industry interaction, and their selection of particular colleges. The objective of this paper is to provide a general ranking and a preferential ranking of top 28 Bschools in India. The approach is to provide accurate rankings by researching on a lot of existing data and by taking students preferences into consideration.

In the era of any information on fingertip or on one click, medical diagnosis is context in which wrong diagnosis should be avoided using extensive information related to patients and symptoms. There should be an efficient system in diagnosis in terms of expert diagnostic opinion within short span of time, so that disease should be prevented to become chronic. To streamline this expert diagnostic opinion process to the patients, in daily routine, Expert System (ES) using artificial neural network can be employed. It is the method which can simulate two very important characteristics of humans, learning and generalization. Using ANN algorithms various types of medical data are handled and output is achieved with defining various relations between that data. Radiology is one of the branches of medical science in which various medical imaging techniques are used to diagnose difference internal medical problems. Digital Image Processing is the science of processing various digital images: such that important information will be generated. An Expert System is also an efficient tool from which diagnosis can be made. Integrating outcomes of neural network from diseased X-ray, to the knowledge based expert system; an expert opinion of diagnosing disease can be generated. In this paper a model is proposed for diagnosing, seven lower lumbar problems as degenerative diseases.

In today’s scenario, there is quick evolution in each field which contains majority and distinctive sorts of information. In order to differentiate sample data from the other, the amalgamation of data mining techniques with other useful algorithms is done. Android development is one of the major arena where there is tremendous need to execute these calculations. Combining frequent pattern calculation with clustering is extremely efficacious for android. In this paper the work is done in two levels, initial stage concentrates on generation of clusters and final stage deals with finding the frequent patterns.

In 1984 Bennett and Brassard proposed a Quantum Key Distribution (QKD) protocol known as BB84 protocol to distribute a random and frequently changed key using quantum mechanism. A major problem in this Protocol is to prove authentication. One of a solution of this problem has already been proposed in [4]. But in presence of Hardware Fault or Interception, above protocol is not applicable. In This paper Proposed System, Key Distillation has been added to overcome Hardware Fault or Interception with minor changes were not available in [4] and was available in original algorithm. It may increase performance of the Proposed System.

Routing in WSN has dependably been a serious issue of concern mainly because of a few case studies which extend from unfriendly deployment conditions, network topology that change over and over, network failures, resource constraints at each sensor hub to issues in designing of routing protocols. Accordingly, the implementation of routing protocol is influenced by a few fundamental elements which must be thought of before any attempt at designed routing are implemented. Two major protocols used in WSN are Dynamic-Source Routing (DSR) and Destination Sequenced Distance Vector Routing (DSDV). DSR protocol is mainly source based routing protocol and implemented to limit the bandwidth utilized by packets in WSN by avoiding the regular messages transmitted to update table in table-driven approach. The proposed system is Multilayer Buffer Management DSR (MBMDSR) where multilayer buffer management mechanism is implemented in existing DSR protocol and a considerable increase in performance was noted.

ZYNQ 7000 Embedded Processing Platform SOC is chips includes ARM dual core A9-MPCore Processor Processing System-(PS-Microprocessor) along with Xilinx Programmable Logic (PL)-Artix 7 FPGA on a single die. ZYNQ SoC provides the high performance and computing throughput at low power using PS along with the flexibility of PL. ZYNQ SoC incorporates independent interfaces for communication of data control signals between PL and PS in various configurations to access the system resources. This paper describes the performance evaluation of such interfaces in terms of resource utilization and power consumption. Here, in this paper, data transfer from PL to PS using low speed AXI GP port and high speed bus like AXI HP port and ACP (Accelerator Coherency) port is discussed. The paper includes design, implementation and testing results on Zynq-7000 SoC based Avnet Zed board.

Recommender systems (RS) try to solve information overload problem by providing the most relevant items to users from a large set of items. Collaborative filtering (CF), a popular approach in building RS, generates recommendations to users based on explicit ratings provided by the community of users. Currently many online platforms allow users to evaluate items based on multiple criteria along with an overall rating instead of single overall rating. Previous research work has shown that considering these multiple criteria ratings for recommendations improved the predictive accuracy of recommender systems.In this paper, we propose a novel approach to increase predictive accuracy of multi-criteria recommender systems (MCRS). Firstly, we use matrix factorization to predict individual criteria ratings and then compute weights of individual criteria ratings through linear regression. Finally we predict overall rating using a weighted function of multiple criteria ratings. Through experiments on Yahoo! Movies dataset, we compare our proposed approach to baseline approaches and demonstrate its effectiveness in terms of predictive accuracy measures.

The solution presented locates locality of reference during the deadlock detection process by migrating detector agents to query multiple blocked agents. To message each blocked agent individually and gather their responses at the shadow agent itself is an alternative to this single migration. The pseudo code provides a context for the solution and insight into the responsibilities and activities performed by each entity.

With the increase in the diversity of data available on the web, excellence of various searches and the need for personalizing the search results arises. The densely distributed web and heterogeneous information environment creates challenges for search engines such as Storage space, crawling speed, computational speed and retrieval of most relevant documents. It becomes difficult to identify the relevancy of the result due to instability in the search query context. In this paper, the framework to personalize web search through modeling user profile by content based analysis and recommendation model is proposed. The framework will use knowledgebase in form of query hierarchy which is specified for individual user to filter discovered results. The proposed approach is also used to discover current search context of particular user by alluding useful links through item-item collaborative filtering techniques. Due to integration of content based analysis and item to item collaborative filtering algorithm, the proposed framework will retrieved the results of user context on query and also suggest links that had been already clicked by the users within same context.

Travel, Tourism and Hospitality industry is growing day by day. It is providing immense opportunities of growth and development to a destination, region, state and Nation. The rapid changes in Information Communication Technologies generated the demand of e-marketing, e-commerce and content marketing in travel, tourism and hospitality industry. No doubts that ICT developments are impacting the demand and growth of tourism and its products phenomenally on the globe. On the other hand the region like Hadoti of Rajasthan state which is highly rich in tourism is lacking behind in the usages of e-commerce-marketing and content marketing in promotion of the attractions and other itineraries. The paper tries to explore the benefits of using e-commerce-marketing and e-business model for the destination image building and promotion of tourism of the region. The paper showcasing the benefits of shifting from the traditional promotional techniques to adapt e-commerce tools and techniques for travel, tourism and hospitality industry of Hadoti region.

As Data Stream Mining is trending topic for Research nowadays and more users increases day by day with online stuff, the size of big data is also getting larger. In traditional data mining extracting knowledge is done mostly using offline phase. While in data stream, Extracting data is from the continuous arriving data or we can say from the online streams. Due to continuously arriving data, it cannot be stored in the memory for processing permanently. So examining of data as fast as possible is important. In this paper we would be interested to discuss about the data stream mining and the issues of stream classification, like Single scan, Load shedding, Memory Space, Class imbalance problem, Concept drift, and possible ways to solve those issues.

One of the most popular approach for personalized recommendations is the Collaborative filtering methods. The key point of this method is to find out similar users by calculating the similarities among them. The product is recommended to a user based on user-user similarity. For finding similarities, the measures like Cosine, Pearson correlation coefficient, Proximity-Impact-Popularity (PIP) measure, and Proximity-Significance-Singularity (PSS) measure can be used. The main issue with the recommendation system is the new user cold start problem where less ratings are available with the user. The user-user similarity matrix obtained with the help of above mentioned measures in terms of cold start problem is not more accurate. In this paper, we show that the hybrid measure is giving more accurate result then the other similarity measures. In this paper a detailed algorithm for the recommendation of a movie to a new cold start user for the hybrid measure is given. The experiments are done with Movie Lens dataset and the results are displayed in the form of user-user similarity metrics in order to obtain the user-user similarity matrix for the hybrid measure.

Self-Organized Wireless Sensor Network (SOWSN) is a system of sensor nodes that takes global decisions through local interactions without involvement of any central entity. Wireless sensor nodes have constrained processing capability and energy. The key characteristic used to evaluate performance of Wireless Sensor Network (WSN) is its lifetime which depends on residual energy of nodes; hence the major challenge in WSN is the efficient use of available energy. Node clustering saves energy and also shows self-organization because global decision like Cluster Head (CH) selection is taken through mutual communication between nodes. In this paper, a new clustering method based on self-organization is implemented to boost lifetime of WSN. Sensor network is divided into regions. Cluster formation relies on Residual Energy (RE) and nearest Distance (D) from CH. Node with highest residual energy becomes CH. Rest nodes join the nearest CH. Clusters are broken when residual energy of CH falls below threshold energy; causing the sensor network to get self-organized into new clusters. RED also focuses to solve the energy hole problem caused due to higher energy consumption by CHs near Sink Node or Base Station (BS).

The crucial part of image segmentation is the proper selection of initial contour to start with the efficient process. The main reason for such kind of segmentation is to reduce the human interaction and moreover to have more accurate results. In this paper we trying to objectify the cyst diagnosis problem in periapical images with the help of automatic segmentation. We have utilized the internal and external energy of image forces that pull it toward features such as lines and edges, confining them precisely. Scale space continuation can be used to develop the catch area encompassing a component.

In this epoch, current web world where peoples groups are associated through correspondence channel and the majority of their information is facilitated on the web associated assets. Thusly the security is the significant concern of this internet community to protect the resources and to ensure the assets and the information facilitated on these networks. In current trends, the greater part of the end client are depending on the end security items, for example, Intrusion detection system, firewall, Anti-viruses etc. In this paper, we propose a machine learning based architecture to distinguish existing and recently developing malware by utilizing network and transport layer traffic features. This paper influences the precision of Semi-supervised learning in identifying new malware classes. We show the adequacy of the framework utilizing genuine network traces. Amid this research, we will execute and design the proactive network security mechanism which will gather the malware traces. Assist those gathered malware traces can be utilized to fortify the signature based discovery mechanism.

The mission of connecting everything on the earth together via internet seems to be impossible. There will be the great effect on human life by Internet of Things (IOT), because with the help of IOT, many impossible things will become possible. IOT devices generates big data having useful, valuable and highly accurate data. It is difficult to extract the required information or data from the set of big data discovered by any device. For this purpose, data mining is used. Data mining will plays important role in constructing smart system that provides convenient services. It is required to extract data and knowledge from the connected things. For this purpose, various data mining techniques are used. Various algorithms such as classification, clustering, association rule mining etc. helps to mine data. This paper represents the different Data mining techniques, challenges, and Data mining issues with IOT.

The following paper propose a new memory management technique of handling the GState’s in DDK and the same can be extended for sending command buffer which hold the job request to various USM/UVC blocks in GPU. We shall discuss the most generic design followed across various device driver and explains the drawback of using the existing design and introduce the new memory management technique along with some futuristic improvements. This design philosophy is more suited in GPU architecture’s which have deferred multi-pass rendering for 3D graphics pipeline.

In this paper, we will present a novel approach for classifying and predicting airline passenger ratings for ancillaries using unsupervised learning techniques like K-Means and Expectation Maximization clustering. The datasets chosen for this study belong to Indian Low-Cost Airlines. The goal is to perform an empirical study and exploratory analysis for predicting the overall rating with respect to the individual ancillary services ratings. Our results suggest that while there is no clear pattern among the ratings that can lead to the overall rating from passengers, the factors like value for money can largely influence the overall rating. Low-cost airlines aggressively promote competitive fares and choice of ancillary services hence the passenger behavior towards the overall rating varies across the airline datasets.

Vehicular Ad-hoc Networks (VANETs) is a growing interest and research area over recent years for it offers enhanced safety and non-safety applications for transportation. Vehicular Ad-Hoc Networks (VANETs) is a developing technology that is yet unclear to many security issues. Security and routing concerns that are unique to VANET present great challenges. In this paper, we have proposed a scheme basically solve the black hole and gray hole attack in vehicular ad hoc networks. Black-hole and gray hole are the well-known routing attacks through which malicious nodes try to downgrade the communication performance of VANETs. In this paper, we focus on recent proposals that aim to enhance VANETs security and routing in systematic and architectural approaches.

In all the modern browsers, maintaining user’s web history is one of the primary tasks. Browser history will help to summarize the activity of the user during a certain period. However, current browser history is not so efficient to visualize in a user-friendly manner and also doesn’t provide enough information for personalized recommendations. One of the key reason is that browsers never maintain any inter-connection between history items. Overall history is maintained in a linear fashion with no information about how the user reached to a particular state. Another issue is that it is not possible to calculate how much time the user spent on any particular website using current history system. This paper provides a conceptual idea of solving these issues by providing a framework that solves this issue by introducing linked data and also describes how this can benefit in improving user experience and quality of recommendations.

While collecting and gathering the data from various domains of a city, one never fail to criticize the city officials as to how they have improved the conditions of livability, economic development and sustainability. A city is always taking initiatives to augment the existing framework or to implement a new framework altogether for transportation system, solid waste management, water supply system and other mission specific domains by gathering relevant data. This process of endless growth of data marts and datasets and analyzing it to gain valuable insight into the domain is being known by the business society today as Big Data Analytics. In our research work, we discussed the challenges and hurdles that are faced during the development of a smart city. Along the way, we will also make use of some of the recent advances in cloud computing, sensing technologies and provide a framework for the smart city using internet of things which can effectively contribute in the enhancement of current models of smart city. We discussed the components of our framework which is proposed for the city, and analytics for large datasets with their applications in various domains.

Multi-focus images may be fused to get the relevant information of a particular scene. Due to the limited depth of field of a convex lens of a camera, some objects in the image may not be focused. These images are fused to get all-in-focus image. This paper proposes an innovative way to fuse multi-focus images. The proposed algorithm calculates weights indicating the sharp regions of input images with the help of Deep Belief Network (DBN) and then fuses input images using weighted superimposition fusion rule. The proposed algorithm is analyzed and examined using various parameters like entropy, mutual information, SSIM, IQI etc.

Big information could be a new technology to spot the dataset’s giant in size and complication. The tremendous growth-rate of huge information with appearance contemporary scientific techniques for informative assortment, large amount of medic’s specialty and Health scientific discipline. Giant amounts of heterogeneous medic’s information became on the market in varied aid organizations. This Medical information may be associate sanctionative resource for account insights for up concern delivery and reducing misuse. The immenseness and complication of that dataset’s yield challenges in analyses and succeeding applications to sensible medical surrounding. There is a sensible problem to judge and infer the info victimization inevitable ways. For extracting helpful data in effective and economical information investigation ways are necessary. Data processing bunch methodology is one that helps in characteristic attention-grabbing patterns from huge information. Several smart applications wide used formula is k-mean. This k-means formula is computationally valuable and conjointly the following cluster quality is heavily depending upon the selection of the initial centurions. This paper proposes a k-means formula is with refined initial centurions. To figure out the initial centurions associated with nursing improved methodology to various clusters the data points is also an assignment. The final results show that the projected formula produces clusters with higher precision in less than working out time.

One of the main challenges in Big data is the processing speed and scalability. Solid State Drive (SSD) helps for faster processing than HDD. Here along with SSD, Spark is also accompanied with hadoop framework for more scalability and fast processing. Apache Spark is a general-purpose engine for large-scale data processing on any cluster. It is a framework which can afford more than 8000 nodes in a cluster Spark allows for code reuse across batch, interactive, and streaming applications. Spark is much faster than MapReduce. It was generally coded from Java; Spark supports not only Java, but also Python and Scala, which is a newer language that contains some attractive properties for manipulating data. Spark runs up to 100 times faster than Hadoop MapReduce in memory and 10 times faster on disk. This paper tries to integrate spark with Hadoop ecosystem along the SSD. It increases the processing speed.

In this paper, singular value decomposition based color image authentication scheme is proposed. Singular values of the host image are modified with that of a secret image. Modifications are optimized to obtain the maximum robustness without losing visual quality. Experiments are done using color images as host and watermark. Simulation results show that the present algorithm is highly resistant to various image processing attacks with significant improvement in perceptibility. The performance is measured objectively in terms of Mean Square Error, Peak Signal to Noise Ratio and Correlation Factor.

The rapid development in the field of electronics, provide a secured environment for the human to live in. This paper introduces a model, “Intelligent Vehicle Surveillance System”, which is designed to reduce the risk involved in losing the vehicles and it also provides notification of occurrence of any accidents, which will reduce the rate of deaths in a very cost efficient manner. This paper introduces a tracking system which passes alert to the owner of the vehicle immediately regarding a theft or an accident of the vehicle with the precise location of the vehicle. There are different methods to identify and continuously track the location of remote vehicle. This proposed system has a single board GPS equipped with GSM and Arduino microcontroller attached in the vehicle. As the vehicle moves its location gets updated via SMS. User can provide real-time control, by sending messages, controlling the vehicle, changing direction as well as on and off. A Software attached would help to read, analyze, process and store the incoming SMS. This system finds its application in real time traffic monitoring. The current system will provide monitoring information from anywhere.

Many studies have derived multiple ways to achieve security in the server and integrating the data in multiple servers by detecting the misbehavior in the server. The data is secured on server using encryption techniques before dividing into fragments before storing on virtual cloud. This study focuses different perspective of storing data on virtual cloud to maintain integrity by storing the fragments of address of data. Hence, the data remains secure and only the address of the data is transmitted when divided in fragments and data is secured with encryption, so it would be difficult for third party to decrypt and access on server. Thus, security level has been increased on cloud platform though data stored on server is more secure and integrity is maintained throughout the cloud platform.

Internet is now being used as content distribution network also. Internet users are interested in specific contents rather than host machines where the content is located. Named Data Networking (NDN) is a step towards future Internet architecture that would be based on named data rather than numerically identified hosts. Many projects are in progress to architect the structure of the future Internet. It is envisaged that NDN would provide functional efficiency with named content applied as the core concept. NDN offers lot of research opportunities to contribute for design of future Internet architecture. In this paper we analyze the modern Internet Protocol (IP) based host- centric architecture and explore the newer scalable and more efficient architecture, based on content-centric approach. Open research issues in various area of NDN are also addressed in the paper.

Employee churn is an unsolicited aftermath of our blooming economy. Attrition may be defined as voluntary or involuntary resignation of a serving employee from an organization. Employee churn can incur a colossal cost to the firm. However, furtherance to prediction and control over attrition can give quality results. Earmarking the risk of attrition, the management can take required steps to retain the high valued talent. Workforce Analytics can be applied to reduce the overall business risk by predicting the employee churn. Predictive Analytics is the field of study that employs statistical analysis, data mining techniques and machine learning to predict the future events with accuracy based on past and current situation. The paper presents a framework for predicting the employee attrition with respect to voluntary termination employing predictive analytics.

Many studies have derived multiple ways to achieve security in the server and integrating the data in multiple servers by detecting the misbehavior in the server. The data is secured on server using encryption techniques before dividing into fragments before storing on virtual cloud. The fragments in which data is divided could be big or small. For encryption technique I had used Reed Solomon codes to encrypt fragmented data. After using encryption they are stored into different servers. Whenever user will challenge for its particular file, Cloud Service Provider will provide that file in original form. If that data chunk from particular file or file is corrupted on any server so using Reed Solomon Code we can repair our corrupted data or file. So, whenever attacker might attack on the server he might get small portion of that file in encrypted form which is not useful for him. He did not get original file from one server because that particular file fragments are dispersed on different servers. This ensures data security to cloud servers, cloud customers important information. It will also provide security from cloud service provider. And if any misbehavior of server is occurred or any file fragment is corrupted can be detected with automated tools and it can be also traced that on which server the file fragment is corrupted. Hence, in any way the data stored on server is more secure on cloud platform.

Cloud computing is like a daily routine now a day. Even though it has numbers of advantages in technical and business view, still there are some challenges there like data storage security, confidentiality and integrity. Main risk in cloud data is about to trust on cloud owner. Encrypted data is not useful for any computational process, so we cannot store as encrypted data. In recommendation system cloud plays very important role. Using homomorphic encryption, we can perform cloud data analyzation. This paper discusses about different homomorphic encryption technique and solution to recommendation system.

The climate changes effortlessly nowadays, prediction of climate is very hard. However, the forecasting mechanism is the vital process. It is also a valuable thing as it is the important part of the human life. Accordingly to the research, the weather forecast of rainfall intensity conducted. The remarkable commitment of this proposal is in the implementation of a hybrid intelligent system data mining technique for solving novel practical problems, Hybrid Intelligent system data mining consists of the combination of Artificial Neural Network and the proper usage of Genetic Algorithm. In this research, Genetic algorithm is utilized the type of inputs, the connection structure between the inputs and the output layers and make the training of neural network more efficient. In ANN, Multi-layer Perceptron (MLP) serves as the center data mining (DM) engine in performing forecast tasks. Back Propagation algorithm used for the trained the neural network. During the training phase of the proposed approach, it gains the optimal values of the connection weights which, in fact, utilized as the part of the testing phase of the MLP. Here, the testing phase is used to bring about the rainfall prediction accuracy. It may be noted that the information/data is used to cover the information from the variables namely temperature, cloud fraction, wind, humidity, and rainfall.

In the present scenario where so many computational problems are being solved using nature inspired optimization algorithms. Human being the most the intelligent creature, deserves to inspire researchers by its method of opinion formation. Steganography is proving to be the ultimate tool of data security. The technique can be improved further by applying optimization. In this paper, a new technique is proposed where optimization is done on the basis of human opinion formation.

Network lifetime is an important parameter to be considered to evaluate wireless sensor network performance. Lifetime is considered in different contexts i.e. some considered network lifetime when first node dies or some percentage of nodes dies or when all the nodes deplete their energy. In WSN, energy conservation becomes important to improve network lifetime. In a network with even distribution of nodes, nodes that lie near the base station deplete their energy faster and hole is created which leads to disconnected network. Incorporating Heterogeneity in the network increases energy efficiency and improves network lifetime. Heterogeneity is included in terms of energy, links and computational resources. Finding optimal number and positions of Heterogeneous nodes to improve network performance is the main concern of this study.

In the recent biometric community security of biometric information has accepted basic significance. Cancelable biometrics is the one technique to do this, which manage before biometric information putting away in the database transforming the biometric information, in a manner that relative minutiae data is not defiled in the transformed template. This paper shows an investigation of few methods of generating cancelable biometric templates. The security level of current biometric framework is upgraded by the structure of fuzzy vault as far as concealing secret key and ensuring the template, by applying a binding strategy on biometric template and cryptographic key. Apply cyclic redundant code(CRC) to distinguish a real polynomial from an arrangement of competitors in light of its straightforwardness. In CRC based fuzzy vault scheme to overcome issue of blend substitution attack two new module is proposed chaff point verifier along with generator. The systems dispense with a blend substitution assault to enhance general security and, subsequently it can distinguish any alteration in vault.

With a specific end goal to excel in teaching-learning the students and the educator must be responsible to each other. The trust built in such an environment will permit the educator to create a conducive teaching-learning environment. A fundamental tool which can be for the above said purpose is the feedback of students on various aspects of teaching-learning. Feedback is ultimate necessity to ensure effective learning. It helps the students to comprehend the subject being examined and gives a clear direction on the most proficient method to enhance their learning. The present paper depicts the analysis of the feedback using Artificial Neural Network (ANN). The feedback of the students in the form of text messages is converted to numerical vectors by considering the positive and negative keywords. Thereafter the ANN is trained using the above said input to predict whether the feedback is positive or negative. We could really enhance student’s accomplishment and achievement all the more viably in an effective way utilizing the aforesaid approach. It paper conveys the advantages and effects of this approach to the students, educators and scholarly establishments.

Mobile phones have taken a crucial part in today’s transferable computer world. Mobile devices are more popular these days because of their small screen size, lower production cost, and portability. Because of their popularity, these devices seem to be a perfect target of harmful malicious attacks like mobile phishing. In this attack, the attackers usually send the fake link via emails, SMS message, messenger, WhatsApp, etc. and ask for some credential data. Mobile phishing is fooling the users to get the sensitive personal information. This paper presents a comprehensive analysis of mobile phishing attacks, their exploitation, some of the recent solutions for phishing detection. Our survey provides a good understanding of the mobile phishing problem and currently available solutions with the future scope to deal with mobile phishing attacks conveniently.

Video Error Concealment is the error hiding technique in videos. In recent years, there is huge requirement of error concealment in video applications such as in video streaming, entertainment, advertisement, media, security, etc. The simulation on MATLAB for the error videos using Block matching algorithm (BMA) has been performed to achieve the concealed videos. From an error video, error frame is detected using Histogram and correlation. This frame is corrected using BMA. First step of BMA is to divide the current frame of a video into macroblock. Second step is to compare each of the macroblocks with a corresponding block and its adjacent neighbors in the frame or previous frame. Third step is to models the movement in a macroblock from one location to another. Last step is to calculate this movement for all the macro blocks that is comprising a frame. This error block is replaced by correct reference block. The quality of the error video and concealed video is measured using PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index Method). An improvement in quality is observed in concealed video.

Delay Tolerant Networks (DTNs) have high message delivery latency, very low message delivery probability, long delay. Delay Tolerant Networks routing schemes faces challenges of dynamically changing network topology. Due to this, routing in delay tolerant network is primary issue to consider. The main objectives of routing in Delay Tolerant Networks is to maximize message delivery ratio. A common method to maximize message delivery ratio is to replicate messages to encountered nodes. One of issues is knowing to how many replicas of a message to create? The work in this paper uses encounter history and current encounter values to dynamically finding number of message replications required. In this paper, a new message replication strategy is proposed which vary the number of message replications. Each node chooses by itself the number of message copies to create for the replication.

Tiny Encryption Algorithm (TEA) is one of the fastest Encryption Algorithms. It is a lightweight cryptographic algorithm with minimal source code. Due to its simple logic in key scheduling TEA has suffered from related key and equivalent key attacks. Therefore a modified key schedule is proposed for TEA. The new key schedule applies Boolean function based SBox to generate different round keys for TEA. The resultant Modified TEA achieves better security than original TEA. The execution time analysis of modified TEA is also presented.

While evaluating the performance of the company in the market, the stock performance of the company plays a vital role in its evaluation. In this paper, the correlation and mean reverting behaviour of various stocks of Banking (Private Banks) from Indian stock market have been examined. Five Private Sector Banks (ten combination/pairs among them) were selected for the study. Along with the correlation test, Augmented Dickey Fuller Test is conducted to test whether the time series follows the mean reverting behaviour. It has been found that three pairs from banking sector were negatively correlated and that high degree of correlation does not necessarily result in mean reversion between two time series.

It is very difficult to detect and monitor the diseases manually and also needs the expertise in the field so that process becomes time consuming. Hence, image processing can be used to detect the diseases and further giving the correct recommendation for the detected disease will be the better solution because only detection of disease will not be the helpful. Disease detection process using image processing involves: Image Acquisition from farmers, image pre-processing and enhancement, edge detection and segmentation, feature extraction, classification of extracted features. The process will not stop here a correct recommendation is very necessary to prevent losses which are faced by the farmers. So designing the recommendation system which gives the best doses for detected disease is the good solution to the farmers.

The most promising objects of a multi dimensional dataset are identified by a skyline query. In case of a higher dimensional, distributed, large dataset undergoing the frequent updates, the response time of skyline queries becomes intolerable. It can be significantly improvised, if a proper execution plan is used for the subsequent queries. In this paper, we have proposed a skyline computation model, SCP. The model presents certain strategies which make use of results of the pre-executed queries. Using these strategies, the execution of the subsequent queries is planned in order to achieve a positive gain in response time of the overall skyline computation. The model is suitable for a distributed dataset which is update intensive.

Question answering learning systems based on AI have been subject of study since long. AI helps in fast access to most suitable answer in a database of question/answer pairs.This paper presents a web learning system (WLS) which answers the frequently asked questions in DSP. The system uses a web Data Store for the answers. The system uses natural language generation (NLG), natural language processing (NLP), question logs and question classification to search out a certain answer to a submitted question. WLS also includes a self-test subsystem.

Rare class classification is important technique in real-world domains like medical diagnosis, bioinformatics, detection of oil spills in satellite images, road accident analysis etc. Imbalanced data classification resides among the top research area of current time and attracted huge interest of researcher. For rare class classification most of researchers have concentrated their study on use of methods like sampling techniques, one-class learning and ensemble based methods. But these methods suffer by some drawbacks. Here we first present an overview of imbalanced data and rare class classification, and then extract unique causes of target class using association rule mining and show importance of causal relationship in determining target class of an instance.

The fourth leading disease in the world today is Diabetes and there are number of challenges to predict and identify the disease. Data mining proposes effective approaches to identify the diabetic patients. This paper proposes clustering and classification of effective diabetes diagnosis based on computational intelligence techniques using PCA with kNN. Diabetes disease data is used to identify feature of clusters. Diabetes disease diagnosis proposes novel distribution function applied to classify each patient. This proposed procedure defines clusters and similarity measure based on classifying with each cluster using computational intelligence techniques. PCA using diabetes disease data for dimensionality reduction. Novel similarity measure is proposed in kNN for classification. Accuracy measures are computed for each patient.

Bigdata comes into big picture in early 2000, since it becomes focus of researchers and data scientist. Main purpose of research and development in the field of Bigdata is to extract and predicts meaningful information from large amount of structured as well as unstructured real world data. In this paper, systematic review of background, existing related technologies used by various big enterprises, data researchers, government officials has been discussed. In addition, presented standardized complex processes to extract useful information such as data generation, storage, modeling/analysis, visualization and interpretation. Finally discusses open issues, challenges and point out the emerging directions in which researchers can work in the age of Bigdata

The dominant degrading factor of quality in ultrasound images is mainly due to the occurrence of speckle noise that in turn leads to false ameliorative decisions, restricts auto diagnosis and telemedicine practices. In medical image analysis speckle reduction is contemplated to be the pre-processing task that sustains decisive information and exclude speckle noise. Meta-heuristics optimization algorithm were used now a days for speckle reduction problems. Our contribution in this paper analyses the use of optimization technique in determining the best noise removing filter coefficients that removes the speckle content contributively. The proposed method comprises the use of Finite Impulse Response filter receiving the filter coefficients from Tree Seed optimization algorithm. Evaluation of noise removal with standard metrics such as Peak Signal to Noise Ratio, Correlation coefficient and Structural Similarity Index shows that the proposed method gives optimal speckle reduction score when compared with conventional filters and its superiority in despeckling medical ultrasound images. Assessment of the proposed methodology with advanced evaluation metrics ensures the ability of it in terms of preserving edges and textural features.

Soil is one of the natural material, which has the different features for the particular characteristics. In digital image processing is the principle to simplify the identification of soil features. Soil consists of both physical and chemical characteristics. These characteristics are used to find the field of soil usage. Thresholding is the conversion of colour image into binary image and that is used for shape based identification. It applicable for feature extract from curvature, valleys, and non-smoothening surfaces and it enhances the feature and get more information. Fractal dimension is one of the soil feature. A new model is proposed to assign various threshold values apply to the same sample and to determine the range and also the best image model (Red-Green-Blue, Hue-Saturation-Value, Hue-Saturation-Luminance and Hue-Saturation-Intensity) of soil samples. The device can also be modelled as most powerful tool for prediction of land usage for various fields such as agriculture and construction.

Recent years have seen an exponential increase in cloud computing for services such as high computational capabilities, vast storage, applications etc. that they provide. Moreover, today’s world is a constant shift from desktop or laptop devices towards handheld smartphones because of the wide range of applications that are supported. This paper discusses the present state of cloud computing for mobile, termed as MCC or Mobile Cloud Computing. This paper also explains the MCC architecture, and the various existing application models, namely Energy and Performance based, constraints based, and Multiple Objectives based models and their examples. We also identify the research gaps in this area which cover the need for standardization of mobile cloud execution platforms. We further point out the vulnerability smartphone clones, and the need to address the privacy issues and security attacks against the user.

MANET consists of collection of mobile nodes with none infrastructure. Each node in Manet organizing yourself and with help of radio links develop a network with the objective of maximizing the mobility into wireless, mobile and autonomous domain. The area, as wherever speedy readying and vigorous reformation are needed and wired network is not obtainable are the popular applications of Manet. Most of the manet applications are depends upon multicast operations. Multicasting is often used to ameliorate the potency of the radio link whereas transmitting numerous messages to escapade integral broadcast identity of radio/wireless transmission. Remarkably the multicast performs high responsibility in MANETs. In Manet, multicast routing should address a large space of problems like low energy, low bandwidth, and mobility. MANET provides subsidiary bandwidth than their wired counterparts; consequently, gathering information about creation of a routing table overpriced and expensive from energy purpose of. Therefore, Energy is one amongst the leading problems in MANETs due to extremely vigorous and dispersed personality of nodes and its superlative use has become an essential demand as all nodes are battery power-driven and therefore the failure of 1 node might have an effect on the entire network. This paper addresses the difficulty of energy optimisation in Mobile Ad hoc networks. The reduction of energy in an exceeding node might increase the likelihood of network ripping. In this paper, we have got tried to scale back the likelihood of such ripping.

Structural brain imaging plays an essential role in acknowledgement of variations that presence in brain relevant to Alzheimer’s disease and different kind of brain disorders. Mostly MR Imaging has preferred because of its higher resolution capabilities to diagnose AD than other modalities. Magnetic resonance imaging is an efficient for visualization and diagnosing various brain disorder in brain. Pathology segmentation for differentiate the diseases affected region to make separation of necrosis and similar damaged tissues cause by disease from normal tissue using clustering principle take up in image processing. Clustering is implemented to make grouping similar characteristics pixels together as a group. In this paper k-mean clustering is performed to separate White Matter (WM), Grey Matter (GM), Cerebrospinal Fluid (CSF), Lateral ventricle, hippocampus region as different individual group with accomplice of Partial Swarm Optimization (PSO) in brain. Different k- mean cluster initialization methods were executed and an exact segmentation were done using PSO k-mean clustering. Volume of both grey matter and white matter are estimated to make comparison with the bench mark images for classifying the various stages of AD.

In software industry, testing effort estimation has always been a challenging task. To keep up with customer expectations and demands, good quality software is a must. For this purpose, a great deal of time and cost is spent on software testing during software development process. To develop an software in an increasingly competitive and complex environment, a lot of techniques or models had been introduced or experimented. Recently, Use Cases for software effort estimation has gained wide popularity. And also there are various modifications like e UCP, Re-UCP etc. that have been used for effort estimation depending upon the type of project. Like e UCP, Re-UCP etc. This paper proposes a new model for fuzzy technique by integrating fuzzy technique and Re-UCP method for reliable effort estimation.

This paper focuses on the Fault Diagnosing methodologies crucial for attaining reliability and maintainability of all electronic circuits is implemented for analog to digital converter (ADC) with a wide range of faults. Fault Diagnosis (FD) is considered as the pattern recognition problem and solved by machine learning theory. Functional test is is needed instead of structural test for testing complex circuits. Fault diagnosis using Fault Dictionary, Neural Networks and Fuzzy logic are enigmatic or inconclusive diagnosis results which have more debug duration and even inaccurate repair actions that exponentially rises service overhead. The effectiveness of these methods are considered, which cover ability in detecting, identifying and localization of faults, the ability of analysing linear and nonlinear circuits, etc. Recent machine learning techniques like support vector machines (SVM) with kernel functions improve the preciseness of functional FD which reduces the product cost through correct repair process. The proposed Multikernel SVM (MKSVM) methodology gives better results than earlier methods as it works with the fundamentals of machine learning and generalization for FD.

Digital watermarking is fragile to different authorization and copyright protection in spatial domain, So in the vast majority of watermarking strategy’s transform domain is utilized. In that paper we have an review of such computerized watermarking procedures and techniques like discrete wavelet transform (DWT), singular value decomposition (SVD), discrete cosine transformation (DCT) with various methodologies alongside their applications, advantages and limitations.

The problem of leaking distribution system is very important issues across the world to operate and via moving steps in this direction better performance of services from water supply organization can be achieved. Even though the methods and technology used in a leakage localization are based on only one kind of sensor, therefore the leakage is not identified until the water has risen above the surface. Due to physical constraint and unique feature of water distribution network designing effectively identification of leakage is very difficult. This paper incorporates idea to propose a new effective practical approach to collect the information from the sensor and after the analysis and computation of that data; information is communicated using any technology like Bluetooth, wireless network, wired network etc. which will be helpful to fire some important decision and based on this valuable decision, leakage control parameter can be controlled using Internet. This paper aims to propose the use of technology of this era, Internet of things (IoT) integrated with recent advances in electronics embedded technology to secure the most valuable resource water for this era as well as for future generation. This paper aims for proposing use of multi-sensor fusion data and Internet of things for leakage detection in pipes for water distribution network.

Holography is the science and practice of making holograms. Holograms currently have a very wide scope for development and is going through various changes at a very fast pace. Due to its multi-faced and multi-dimensional nature many tech pioneer companies are trying to adopt and develop it for various aspects. Since holograms are 3-Dimensional images or stream of images it creates an illusion of depth for the content it is displaying. With holograms one can create rich and immersive content which takes the user experience to a whole new level. This proposed system can be used by common people to view images and videos as mid-air holograms at home. The new implementation includes interaction with the holograms using gestures via leap motion technology. The users could use hand gestures to change, rotate, zoom in or zoom out the images or play, pause, go forward or go backward in a video.

In recent years, the smartphone has become a powerful miniature computer owing to its computing power, portability, and flexibility. The integration of smartphone and cloud computing technology offers significant benefits and expand the capabilities of smartphones to support applications that demand high computational resources. However, the proliferation of popular technologies increases potential opportunity for misuse and illegal activities. Thus, the advancement of forensic research is inevitable to cope with the recent developments in mobile cloud platforms. The adoption of forensic methodologies to the mobile cloud investigation is still in its infancy stage. Mobile Cloud Computing (MCC) offers unique challenges compared to traditional digital forensics due to its ubiquitous nature. In this research work authors have emphasized on various solutions available for mobile cloud forensics. Authors have analyzed and discussed the scope and possibilities regarding legal implications towards forensic analysis in MCC, potential challenges, and its countermeasures.

Medical Science has grown widely with the advancement in technology and research. The retrieval of accurate information regarding medical terminologies has become a major requirement for progress. Everyone will need to access this information starting from layman to the expert doctors. Each one will have their set of terms of accessing same information, therefore certain techniques and methods are required to retrieve precise information. World Wide Web reduces the task of generalizing all information to a single platform. Hence the task remaining is to map the query of the user to the appropriate concept in medical science. This paper surveys on existing methods and tools that help to retrieve accurate information as per the query of the user belonging to any knowledge group.

The increase in GDP of the country has given a flight to industrialization and urbanization, causing more and more utilization of agricultural lands for non-agricultural purposes. Since the availability of agricultural lands is limited, requisite measures must be taken to restore wastelands for cultivation. Therefore to filter out the suitable wastelands for reclamation and predict their level of utilization, this paper proposes the agricultural intelligence decision support system. The proposed system has two phases. The first phase consists of the mining technique in which required attributes are selected, intersection is applied as spatial predicate and weights are assigned to linguistic terms for obtaining weighted fuzzy rules. In the second phase the fuzzy inference system is constructed in accord of the weighted fuzzy spatial rules mined in the previous phase. This will assist agriculture-related organizations and persons to take well informed decisions for effective utilization of wastelands.

Role of Geographical Information is significant in agriculture and it can also play an important role in boosting agriculture based e-commerce. With majority of products available on internet, agricultural products still behind to be online. Lacking point behind this situation is the improper application of Geographical Information, as it would act as the backbone in this value chain. In spite of various steps are taken to acknowledge e-commerce related activities in the field of agriculture profit does not percolate to farmers. This can be accelerate using Geographical Information into the current stream of e-commerce. In majority of the scenario merchant act as intermediate which pilfers the seller’s and buyer’s share. This paper presents a merchant free platform Geographical Information Assisted e-Commerce (GIAE) which uses Geographical Information of the user and allow selling and purchasing based on rank of either party, so that both can gain maximum benefit.

Digital Watermarking has been discussed as a potential solution against illegitimate use of videos over the globe. In this paper, blind hybrid video watermarking technique in transformation domain using differential embedding approach is analyzed. Discrete wavelet transforms and discrete cosine transform with differential embedding is used. The embedding and extraction of watermark is described and watermarks are examined in terms of invisibility and robustness respectively. The method performance in terms of peak signal to noise ratio and bit correction ratio is analyzed. The numerical results show the method has better performance when compared with DWT alone and DCT alone methods under noise and different video processing operations.

To help specially abled learners few Intelligent Tutoring Systems solutions exist. However, these tutoring systems lack in construction of individualized dynamic sequences and on the go learning content provision as per the learning ability of the learner. In order to overcome this limitation, there is a need of an intelligent system with a distributed approach that will handle this issue. This paper describes an architecture using Multi Agent Simulation model that where the agents interact with each other to cater to the individual needs of the specially abled learner. The paper suggests a simulation model where the agents such as Pedagogical agents, Intelligent Sequencing agent and Feedback with their detailed design.

Social media proved to be a medium to document our daily lives. It is a virtual tool used by an individual or organizations for various communications and information exchanges. Data available on social media assists in solving crimes; like to identify people, locations, evidence gathering notifying public, community, soliciting crimes etc. It is desired to attain privacy & security on these platforms. Further, it is observed that the information exchange and forwards are persuasive messages. These messages are intended to attain economical, political, religious, regional gains. The study in this paper proposes a model to analysis persuasive messages on social media. The results of the analysis demonstrate that these messages are the potential contents which leads to cyberpsycho attacks. The work in this paper is a step towards a study of threat analysis which demonstrates the psychological health of society & cyberpsycho attacks.

Optimal forecasting of ATM cash repository in an optimal way is a complex task. This paper deals with cash demand forecasting of NN5 time series data using neural networks. NN5 reduced Dataset is a subsample of 11 time series of complete dataset of 111 daily time series drawn from homogeneous population of empirical cash demand time series. Main objective of this paper is to forecast cash demand forecasting of NN5 data with neural networks. Further, the same process is applied on clusters of ATMs. Discrete time wrapping is used as distance measure. Root mean square error has been calculated for such clustered group of ATMs and average is calculated. Root Mean Square error indicates applications of clustering before applying Neural Network increases precision in forecasting of ATM Cash Repository.

This paper proposes a methodology to compute trust and rank user on the basis of time, number and nature of interactions. Technique models trust based on two main factors: Engagement of user, temporal/behavioural factor based on difference between time of user reaction to that of any social media activity. It is evaluated using verified Facebook page of ‘Panjab University’ with 160,000+ users. Validation uses 985 active users. On analyzing the result it is found that the top scorers are from ‘Public Relation Department’ of university. Naïve Bayesian machine learning technique has classified the data more accurately as compared to SVM and Logistic regression. Accuracy of Naïve Bayesian is 77.54%, as the condition of independence of dataset is satisfied and degree of overlapping is null. This result states that proposed model provide efficient technique to rank users on social networks.

Automatic Person Recognition and Authentication faces a huge challenge especially in high risk environment. Multimodal Biometrics do address this issue up to a certain extent. Voice is the simplest and distinct single modality and lip movement is unique for each speaker. Dynamic movement of the lips if tracked online while a Speaker is speaking can authenticate a speaker in real time and the imposter can be identified thus making the biometric system robust and secure. This paper detects the lip movement of a Speaker, creates contours around the mouth region in real time and creates data base for a Speaker. Using the USB webcam, face detection was done from which the mouth region was segmented and contours were created around the lip area with the help of random points selected on both the upper and lower lips which were mathematically obtained. Based on these points of movement, a database was created for an individual. The time taken to create the speaker’s lip contour database takes hardly a few seconds which is the advantage and requirement for Real time applications which we have achieved.

In digital cameras, optical lens suffers from a limited depth of focus; therefore, it is limited in providing sufficient information regarding the scene. Two images with different depth of focus can be used to produce an “all-in-focus” image in order to provide better description. The concept of merging two different multi-focus images having different focus content using the combination of Wavelet and Principal Component Analysis (PCA) is carried out in this paper. In the proposed multi-focus image fusion method, each of the registered source images are decomposed into approximation and detailed layers using Discrete Wavelet Transform (DWT). Later, average fusion rule is applied on approximation layer and PCA is applied on detail layer to preserve both spectral as well as spatial information. Image quality assessment of the fused image is made using Standard Deviation (SD), Fusion Factor (FF) and Entropy (E) which justifies effectiveness of proposed method.