Proceedings of ICRIC 2019

Absenteeism has become a severe problem for many organizations. The problem posed in this paper was to build a predictive model to predict the absenteeism for MNCs by previously recorded data sets. This exercise not only leads to prevent or lower absenteeism but forecast future workforce requirements and suggests ways to meet those demands. For faster processing of massive data set, the data was analyzed efficiently so that we get the minimum response time and turn-around time, which is only possible when we use the right set of algorithms and by hard wiring of the program. Different machine learning algorithms are used in the paper that includes linear regression and support vector regression. By analyzing the results of each technique, we come across that the age parameter mainly affects the absenteeism that is linearly related to absenteeism.

IoT helps the society to solve various problems in medicine. There are many problems, which can be solved using IOT healthcare such as detection of diabetes, detection of diabetic foot ulcer, abnormality in heart rate and many more. This paper presents the design and implementation of IoT-based system to be used in healthcare for detection of diabetic foot ulcer. The model will monitor the health of diabetic foot ulcer patient and will send alerts if found any abnormality. The development of this model will be done on Node MCU development board. This model enables users to record, analyze and send large data to the users in real time and efficiently. This will help in reducing visits to doctors and will help in live alerts and the abnormality in the patient.

IoT-based transportation system is getting smarter and smarter to provide quick, safe and reliable services to the user. This smarter transportation system is called Intelligent Transportation System (ITS). ITS incorporates wired and wireless communication, electronic technologies, computational technologies, cloud platforms, GPS and sensor to assist user to be informed on road safety and make safer, coordinated, comfort and ‘smarter’ use of transportation medium. ITS is an advanced IOT application that connects huge number of objects to communicate with each other. As number of objects connected to ITS application increases, we face with a challenge of adding value to raw sensor data. The focus of this paper is to address this challenge with a context—aware model. Also, the effectiveness of context—aware in ITS is illustrated by discussing different real time scenarios.

Android is one of the smartest and advanced operating systems in the mobile phone market in the current era. The number of smartphone users based on the Android platform is rising swiftly which increases its popularity all over the world. The rising fame of this technology attracts everyone toward it and invites more number of hackers in Android platform. These hackers spread malicious application in the market and lead to the high chance of data leakage, financial loss and other damages. Therefore, malware detection techniques should be implemented to detect the malware smartly. Different techniques have been proposed using permission-based or system call-based approaches. In this paper, a hybrid approach of static, dynamic and intrinsic features based malware detection using k-nearest neighbors (k-NN) and logistic regression machine learning algorithms. The intrinsic feature contribution has also been evaluated. Furthermore, linear discriminant analysis technique has been implemented to evaluate the impact on the detection rate. The calculation uses a publicly available dataset of Androtrack. Based on the estimation results, both the k-nearest neighbors (k-NN) and logistic regression classifiers produced accuracy of 97.5%.

In today’s world, machine learning has gained much popularity, and its algorithms are employed in every field such as pattern recognition, object detection, text interpretation and different research areas. Machine learning, a part of AI (artificial intelligence), is used in the designing of algorithms based on the recent trends of data. This paper aims at introducing the algorithms of machine learning, its principles and highlighting the advantages and disadvantages in this field. It also focuses on the advancements that have been carried out so that the current researchers can be benefitted out of it. Based on artificial intelligence, many techniques have been developed such as perceptron-based techniques and logic-based techniques and also in statistics, instance-based techniques and Bayesian networks. So, overall this paper produces the work done by the authors in the area of machine learning and its applications and to draw attention towards the scholars who are working in this field.

Osteoporosis, a pathological disorder of bones affects millions of individuals worldwide and is the most common disease of bones after arthritis. It is caused due to a decrease in mineral density of bones leading to pain, morbidity, fractures and even mortality in some cases. It is diagnosed with DXA, but its high-cost, low-availability and inconsistent BMD measurements do not make it a promising tool for diagnosis of osteoporosis. The computer-aided diagnosis has improved the diagnostics to a large extent. Deep learning-based artificial neural networks have shown state-of-the-art results in the diagnostic field leading to an accurate diagnosis of the disease. This paper reviews the major neural network architectures used for diagnosis of osteoporosis. We reviewed the neural network architectures based on the questionnaires and the deep neural architectures based on image data implemented for diagnosis of osteoporosis and have summarized the future directions which could help in better diagnosis and prognosis of osteoporosis.

Drug target interaction is one of the most significant fields of research for drug discovery. The laboratory experiments conducted to identify the drug target interactions are tedious, delayed, and costly. Hence, there is an urgent need to develop highly efficient computational methods for identifying potential drug target interactions that can limit the search space of these laboratory experiments. The existing computational techniques for drug target interaction have been broadly classified into similarity-based methods and feature-based methods. In this paper, a novel feature-based technique to predict drug target interactions has been proposed. The technique uses ensemble learning to determine drug target interactions. Ensemble learning offers greater accuracy in comparison with the traditional classifiers. Thus, the proposed technique aims to improve accuracy using ensemble learning. Also, dimensionality reduction of drug target features is performed using principal component analysis so that the computational time of the method can be reduced. The results indicate an improved performance in comparison with the state-of-the-art methods in the field.

In the past decade, the evolution of computing has moved from distributed, parallel, grid, cloud, and now to fog computing. The massive amount of data generated by Internet of Things (IoT) devices is growing up exponentially. The flood of information (generated by those IoT/Internet-connected devices) becomes troublesome for data processing and analytical prediction functionality using cloud computation. Several problems have been investigated with cloud computing with respect to latency, limited bandwidth, low Internet connectivity, etc. Here, solution to such problems can be solved by introducing fog computing with powerful functionality of cloud framework, i.e., based on the deployment of fog nodes called microclouds at nearest edge of data sources. Fog computing for big data/IoT data analytics is in evolving phase and requires extensive research to produce more knowledge and smart decisions. This article discusses several basic facts related to fog commuting, challenges in fog computing and opportunities in the near future, in the context of fog big IoT data analytics. In addition, this work also emphasizes the key characteristics in some proposed research works, those make the fog computing as a suitable (useful) platform for new proliferating IoT devices, services, and applications. Most significant fog applications (e.g. healthcare monitoring, smart cities, connected vehicles, and smart grid, etc.) will be discussed here to create a well-organized green and quantum computing paradigm to support the next generation of IoT applications.

Health care is an emerging industry with all our lives dependent on it. With the advancement of technology, the health care industry is also advancing. With better prevention, diagnosis and treatment options, technology is rapidly integrating itself with medical sciences for the betterment of humankind. However, it is time to delve deeper into this integration. One upcoming technical aspect known as machine learning is a very useful tool when it comes to its application in the health care industry. Machine learning algorithms such as support vector machines and artificial neural networks when combined with the existing medical infrastructure prove not only to be at par with the state-of-the-art technology but also prove to be more efficient and faster than them. This paper looks at the possible applications as well as the current progress of the integration of machine learning algorithms in the health care industry.

Internet of things (IoT) is the hot point in the Internet field. The concepts help to intercommunicate physical objects furnished with sensing, actuating, computing power and hence connect to Internet. With the help of sensor, actuators and embedded microcontrollers, the verdict of smart object is realized. Wherein these smart objects colligate data from the environment of development, process them, and take reasonable actions. Thus, the IoT may generate unbelievable benefits and helps human beings in living a smart and luxurious life. Due to the potential utilizations of Internet of things (IoT), it has ended up being an unmistakable subject of logical research. The significance and the utility of these advances are in sizzling exchange and research, yet on the field of agribusiness and ranger service, it is very less. In this paper, utilizations of IoT on farming and silviculture has been well perused and broke down; additionally, this paper briefly presented the innovation IoT, agribusiness IoT, rundown of some potential applications areas where IoT is exercisable in the horticulture part, advantages of IoT in farming, and displays a survey of some literature survey.

In this epoch of data surge, big data is one of the significant areas of research being widely pondered over by computer science research community, and Hadoop is the broadly used tool to store and process it. Hadoop is fabricated to work effectively for the clusters having homogeneous environment but when the cluster environment is heterogeneous then its performance decreases which result in various challenges surfacing in the areas like query execution time, data movement cost, selection of best Cluster and Racks for data placement, preserving privacy, load distribution: imbalance in input splits, computations, partition sizes and heterogeneous hardware, and scheduling. The epicenter of Hadoop is scheduling and all incoming jobs are multiplexed on existing resources by the schedulers. Enhancing the performance of schedulers in Hadoop is very vigorous. Keeping this idea in mind as inspiration, this paper introduces the concept of big data, market share of popular vendors for big data, various tools in Hadoop ecosystem and emphasizing to study various scheduling algorithms for MapReduce model in Hadoop and make a comparison based on varied parameters.

Over the past few decades, the artificial intelligence is being employed in diverse fields like pattern classification, image processing, object identification, recommender systems, speech recognition, etc. Machine learning has made it possible to develop intelligent systems through training that equip machines to handle different tasks, exactly on the analogy similar to humans. In medical field, machine learning algorithms are being used for prediction, early detection and prognosis of various diseases. These algorithms suffer a certain threshold due to their inability to handle large amount of data. Deep learning based techniques are emerging as efficient tools and can easily overcome the above difficulties in processing data related to medical imaging that includes mammographs, CT scans, MRIs and histopathology slide images. Deep learning has already achieved greater accuracy in early detection, diagnosis and prognosis of various diseases especially in cancer. Dysplasia is considered to be a pathway that leads to cancer. So, in order to diagnose oral cancer at its early stage, it is highly recommended to firstly detect dysplastic cells in the oral epithelial squamous layer. In our research work, we have proposed a deep learning based framework (convolutional neural network) to classify images of dysplastic cells from oral squamous epithelium layer. The proposed framework has classified the images of dysplastic cells into four different classes, namely normal cells, mild dysplastic cells, moderate dysplastic cells and severe dysplastic cells. The dataset undertaken for analysis consists of 2557 images of epithelial squamous cells of the oral cavity taken from 52 patients. Results show that on training the proposed framework gave an accuracy of 94.6% whereas, in testing it gave an accuracy of 90.22%. The results produced by our framework has also been tested and validated by comparing the manual results recorded by the medical experts working in this area.

Smart agriculture with IoT (Internet of things) gives rise to agribusiness and may fulfill the demand of food in the world. Agribusiness will become more productive when IoT is introduced. As the farmers are becoming more aware to IoT, the possibility of introducing new technologies is gaining momentum in agribusiness. The IoT is highly efficient, accessible, persistent and of exposed nature. The farmers may use sensors for monitoring the crops, soil and also analyze the crop production. This will give boost up to precision agriculture. The information about farms like temperature, moisture level and soil mineral level, pH value of soil and air quality can be collected remotely by using IoT on smartphones or computer systems. Today’s IoT techniques and devices are so advanced and are mainly application dependent providing smart systems to be deployed in specific areas. In this paper, a review of IoT devices and protocols is presented and it also throws light on issues arise during the implementation of IoT in agriculture.

Supervised machine learning is used to solve the binary classification problem on four datasets of European Survey of Schools: Information and Communication Technology (ICT) in Education (known as ESSIE) which is supported by European Union (EU). To predict the gender of the principal based on their response for the ICT questionnaire, the authors applied four supervised machine learning algorithms (sequential minimal optimization (SMO), multilayer perception (ANN), random forest (RF), and logistic regression (LR) on ISCED-1, ISCED-2, ISCED-3A, and ISCED-3B level of schools. The survey was conducted by the European Union in the academic year 2011–2012. The datasets have total 2933 instances\ & 164 attributes considered for the ISCED-1 level, 2914 instances\ & 164 attributes for the ISCED-2 level, 2203 instances\ & 164 attributes for the ISCED-3A level and 1820 instances\ & 164 attributes for the ISCED-3B level. On the one hand, SMO classifier outperformed others at ISCED-3A level and on the other hand, LR outperformed others at ISCED-1, ISCED-2, and ISCED-3B. Further, real-time prediction and automatic process of the datasets are done by introducing the concepts of the web server. The server communicates with the European Union web server and displays the results in the form of web application. This smart approach saves the data process and interaction time of humans as well as represents the processed data of the Weka efficiently.

Hard computing, i.e., conventional computing, requires an exactly expressed analytical model. At the time of implementation of hard computing techniques, each time exact results are guaranteed. The fundamental premises and standards behind hard computing are precision, certainty, and rigor. The main problem of hard computing is the time consumption of different algorithms such as kriging interpolation algorithm, Smith–Waterman algorithm, LZW compression, and many more. GPU provides an efficient way to do massively parallel operations such as taking the square root of all values in a region of memory quickly. The operational frequency of GPU is slower than CPU but has more cores than the CPU. This is the main reason why the above algorithms perform better on GPU architecture compared to the CPU. CUDA-based implementation of several hard computing techniques on GPU gives more significant results in terms of time. This article provides an overview of the current literature of CUDA and GPU-based implementation of several hard computing techniques.

In today’s world of the twenty-first century, the Internet of Things (IoT) has emerged exponentially. Many applications are developed in these fields of automation. When it comes to home automation, this concept can be integrated to make it smarter. It makes it easier to access and monitor different home appliances. This paper shows how IoT can be used for smart home automation using NodeMCU and android mobile application. The main focus of the paper is, how the sensor nodes collect the data and pass it to the mobile devices to perform some action as per the user’s commands and provide support to IoT-based mutual controlling between fan and AC based on threshold temperature value.

Real-world datasets in many domains like medical, intrusion detection, fraud transactions and bioinformatics are highly imbalanced. In classification problems, imbalanced datasets negatively affect the accuracy of class predictions. This skewness can be handled either by oversampling minority class examples or by undersampling majority class. In this work, popular methods of both categories have been evaluated for their capability of improving the imbalanced ratio of five highly imbalanced datasets from different application domains. Effect of balancing on classification results has been also investigated. It has been observed that adaptive synthetic oversampling approach can best improve the imbalance ratio as well as classification results. However, undersampling approaches gave better overall performance on all datasets.

Mental health is considered as one of the most sensitive topics of research and it is highly affected by an individual’s mood and sentiments. Social media has been proven to be one of the major catalysts in deterioration and fickleness of one’s mind. In this paper, we present an android application called “moody buddy” ingratiated with a heartbeat analyzing hardware kit which would detect and analyze the moods and emotions of an individual very close to accuracy. Mood recognition and sentiment analysis is a vast and complex area of research. Moreover, monitoring human emotions is found out to be one of the technically challenging aspects. So, in order to achieve the quality output of our research and testing work, we have taken help from artificial intelligence and Internet of Things domain. Here, we have considered the activity of the user on his/her social networking as a starting point of our research work. The concept of logistic regression is used in our software. In order to solidify our idea more, we are adding a hardware component which would monitor the heartbeat of the person and its modulation. In case of any abnormality examined in the heart rate, the questionnaire appears again. At the end, a cumulative of the hardware component’s results and software component’s would help us analyze and detect the current mood of the individual to very close to high accuracy value.

Human liver is believed to be the largest gland in human body. Weight of healthy human livers is around 1.2–1.5 kg and contributes approximately 3.3% of total body weight. Liver breaks down nutrients from our daily diet into substances which are less toxic to our body such as ammonia to a much less toxic substance called urea. According to data released by National Institute of Nutrition (NIN), Hyderabad suggests that food we consume are lesser nutritious then we used to consume in last three decades. According to data published by WHO in 2017, deaths due to liver disease reached 259,749 or 2.95% of total deaths which makes liver disease one of leading cause of death in India. With the power of machine learning and data science, we can provide better information to doctors so that they can start treatment of disease at its initial stage of disease. This paper investigates the performance of logistic regression, K-nearest neighbor algorithm, decision tree and support vector machine algorithm on liver reports of Indian patients. Dataset of Indian patients is collected from UCI repository. Some patient whose age exceeds 79 years is listed as of age 80 years. Algorithms are evaluated on the basis of: (i) recall, (ii) precision, (iii) F1-score and (iv) support.

Internet of things “IoT” is now an integral part of our life. Its widespread usage is almost in every application starting from smart home, smart cities, smart farming, remote monitoring, industrial automation, and transport and the list is cumulative. Although this is very useful technology, many challenges are hampering its growth. The crucial challenges faced in the development of IoT are: security, privacy, constrained resources (as first layer in IoT is of sensors and they are quite constrained as far as resources are concerned), interoperability, and integration. Tremendous work has been done in these directions but much more is needed to overcome these barriers. Based on the literature review, this paper provides comprehensive information on IoT, its applications, approaches, and various challenges faced by existing approaches used in smart devices, which in turn can be helpful in finding future prospects.

Web search personalization is the process of providing personalized results to the user for his query. In this paper, we present a relevance model to personalize search results which is based on query personalization. The user query is directly matched to the keywords of the user profile, and the original query is altered according to the keywords which is more likely similar or related according to the similarity measure. By finding the similarity between the user original query and user profile, a linear combination of preference space is generated at run-time to determine more accurately which pages are truly the most important with respect to the modified query. A heuristic algorithm is used to maintain the user profile based on the ongoing behavior. Our experiments prove that retrieving the search results based on query modification is effective in providing the personalized results to the user.

Wireless sensor networks (WSNs) are collection of sensor nodes. The main goal of wireless sensor nodes is to sense the medium and process the data to specified location. WSNs suffer with many technical challenges such as node deployment, battery power, self-configuration, slow convergence, and packets drop during the data transmission. This paper focuses on efficient data transmission, quick convergence and optimizes lifetime of WSNs using hybrid artificial bee colony with salp (HABCS). This paper also generates the different frame formats such as hello message, bond message, route request message, route reply message, and data packets. The proposed approach is implemented using network simulator-2 and it is more energy-efficient than other existing classical and SI protocols.

The energy efficiency and lifetime of wireless sensor networks (WSNs) are the more focusing points. The WSNs faced many challenges during the data transmission. Node deployment, leader selection, and optimal route selection are challenges that affect the energy level and lifetime of WSNs. Many existing techniques have been proposed to node deployment, cluster leader and optimal route selection. But, all existing techniques have not given satisfactory results in the network energy optimization. Therefore, this paper presents hybrid artificial grasshopper optimization algorithm (HAGOA). It is an inherited behavior of artificial grasshopper optimization and artificial bee colony variance. The proposed algorithm will place sensor nodes using artificial grasshopper optimization technique. These sensor nodes may be static or dynamic that depends on the network scenario. The cluster head selection and optimal route selection will perform using artificial bee colony variance. It also performs balancing between exploration and exploitation phases in the given search space. This algorithm is a combination of two families: Artificial grasshopper and ABC Variance, respectively. It compares with existing classical and swarm intelligence (SI) protocols in the terms of remaining energy, sensor node lifetime, consumed energy, end to end delay and maximum number of rounds.

With the advancement in the wireless communication, there has been an immense growth in the number of vehicles on the road. The unpredictable nature of vehicular ad hoc network (VANET) due to random increase and decrease of the nodes/users (vehicle) on the roads is a challenging issue. Moreover, the increase in the number of nodes creates the problem of spectrum scarcity due to shortage of licensed spectrum for vehicular services. In order to solve the issue of spectrum scarcity, the cognitive radio network (CRN) has been developed which exploits the unlicensed spectrum for communication without affecting the licensed communication that is using interference avoidance. The CRNs are more vulnerable to the security and the privacy of the networks because the transmission parameters required for communication avoid the interference with the licensed and unlicensed users. Moreover, the safety message among the vehicles ensures the safety of the vehicles in the cognitive radio vehicular ad hoc network (CR–VANET) and also manages the sharing of the licensed/primary and the unlicensed/secondary users in the network. The routing and network topologies are a challenging issue due to mobility of vehicles. Therefore, in this review paper, we present the applications and various routing topologies for CR–VANET.

Because of developing concern to environmental changes, renewable energy is being looked as a key alternative to the conventional sources. Photovoltaic power is a green and an abundant renewable energy source. Photovoltaic power is dependent upon the solar irradiation which is highly intermittent in nature. So the precise forecasting of PV power is necessary to improve the operation of an electrical grid with the distributed energy resources. This paper proposes a novel hybrid model by combining particle swarm optimization (PSO) and the feed-forward neural network (FFNN) together. Proposed hybrid model is applied to forecast the PV power in Indian electricity market. One-year data consisting of hourly PV power generation, direct radiation, diffused radiation, and ambient temperature from the Indian energy market has been used for PV power forecasting. The developed model is applied for one-week ahead PV power forecasting for winter, summer, rainy, and autumn season, respectively. The performance of the proposed hybrid model outmatches and compared with some recently reported model.

A type of ad hoc network formed among moving vehicles that come in one another’s radio transmission range is called Vehicular Ad Hoc Network (VANET). VANET may be deployed using three architectures: pure ad hoc, WLAN and hybrid. Vehicle to Vehicle (V2V) and Vehicle to Infrastructure (V2I) are the two modes in VANET used for communication among vehicles. These features not only distinguish a VANET from other ad hoc networks but also make these networks more exposed to attacks and increase their complexity. These networks being the primary mechanism for communication in VANETs, appropriate and timely delivery of information is of prime importance. Due to the existence of various vehicular traffic scenarios, a single category of routing protocols is not sufficient for the VANETs. Therefore, various categories of routing protocols have been tailored to meet specific kinds of routing requirements in this framework. Most significant ones are ad hoc/topology-based, position-based, geocast-based, cluster-based, broadcast-based protocols. There exist a number of attacks that apart from affecting various other parameters also affect the routing protocols in these VANETs. Most of these attacks may be launched in all the above-mentioned three architectures. But, one of the most dangerous attacks is the Sybil attack that may be initiated in pure ad hoc deployment of VANET where vehicles communicate with one another in one to one manner using carry forward approach. In this paper, with the help of illustrative example for each of the five categories of routing protocols, we show how Sybil attack affects these protocols. Considering the ad hoc scenarios in VANETs, we also discuss a prevention mechanism for the Sybil attack briefly.

In this paper, WDM optical OFDM system model with coherent detection has been proposed for different channel spacing. Coherent optical OFDM system is mostly used in various applications because of its various advantages such as high spectral efficiency and flexibility. Four different frequency channels with same power are used and their effects on the output in term of Q factor, minimum BER is analyzed using optisystem software. The proposed OFDM system provide high Q-factor with minimum bit error rate (BER) making system more efficient. At a channel spacing of 100 GHz the system provides the best result.

In the proposed model, effect of number of transmitters and receivers on performance of proposed free-space optical (FSO) communication system is simulated as well as analyzed under various weather conditions like clear, haze and fog. FSO communication provides a strong and efficient method for transmission and reception of information through free channels because of its higher data transfer capacity and inbuilt ability of security. The effect of environmental parameters has been investigated for range of 1 km. The effect of attenuation increases in bad weather conditions affecting the performance of FSO system while designing a system the weather conditions have to be taken care. The effect is to beat the impact of fog attenuation on FSO system by assessment and execution through simulated results of the executed model with respect to Q factor, height of eye, power received and bit error rate. The proposed multiple TX/RX FSO system shows better results as compared to the 1 TX/1 RX system.

Nowadays Wireless Sensor Network (WSN) is popular research field for new scholars. This technique has huge potential in various fields like energy efficiency, data gathering, network security, etc. The bitty sensor nodes are formed a wireless sensor network which has capabilities of communication, transmission and sensing with certain limitations. Limited capability of energy or battery life which attracts many scholars to do the research and find appropriate solutions. Cluster-based heterogeneous WSN (HWSN) is a possible solution to improve the network life time. Keeping in the view of above, this is the key area we have taken in this paper to improve the network life time with the use of less battery consumption. Here, we proposed “Dynamic Distance based Lifetime Enhancement scheme for Heterogeneous WSN”. In the proposed scheme cluster head is selected by dynamic hopping for data transmission to base station. By the selection of dynamics hopping transmission, network life time improved significantly as compared to pre-existing techniques like LEACH, SEECP protocol.

With the passage of time, Wireless Sensor Networks (WSNs) gain importance and have become one of the most fascinating areas of research in the past few years. Reliable and efficient data routing, that helps the data to reach its destination, remains the bottom line of research problem. Hence, various routing protocols are developed which are based on different parameters. But opportunistic routing has fascinated many researchers these days because of its broadcasting nature that makes it more efficient than the old routing methods. However, while routing in this sensor network the loss of data takes place because of its security lack. Therefore, security is also a challenging issue with the routing of WSN. This paper presents an overview of the WSN, its architecture, applications of WSN in different fields presented. An overview of the security aspect of routing in WSN is given. Finally, a comparison of different security methods of opportunistic routing is also presented.

Lately, there has been an increase in the trend among Mobile Ad hoc NETwork (MANET) researchers to use Named Data Networking (NDN) as a network stack solution in MANET. Thus, this paper presents an in-depth review of the potential uses of NDN in MANET environment. In addition, this paper also provides future research direction that could be undertaken on the subject.

This paper seeks to extol the virtues of named data networking (NDN), as an alternative to host-centric networking (HCN), for its prominent features that can be taken advantage of to significantly reduce energy consumption demands in a mobile ad hoc network (MANET) implementation. Therefore, a NDN-based content routing mechanism was compared with two types of HCN routing protocol implementations in this study: OLSR and Batman-adv. The experimental results obtained from this research provide early evidence that NDN can increase the energy efficiency of MANET compared to the use of HCN primarily TCP/IP on the network stack solution for MANET. Of particular note would be NDN-based content routing’s viability as a solution for energy consumption issues that plague wireless multi-hop ad hoc networks. Last but not least, this paper also provides the future research direction that could be undertaken on the subject.

Biometrics is one of the most promising technologies for providing secure authentication in modern computing applications and eradicates the issues associated with the traditional authentication systems. Almost 50% of the security infrastructure comprises fingerprint biometric. As the market share of fingerprints is increasing tremendously, its security is becoming a challenge for research community. In this paper, a brief review of different fingerprint template security schemes has been presented. Moreover, various masquerade attacks on fingerprint template have been studied and their countermeasures are presented. A comparative analysis of different template security schemes based on different performance metrics like FAR, FRR, and EER is also provided. It was seen that the methods employed for fingerprint may not work for other biometric traits like iris, face, etc., because of their difference in dimensions of templates. This paper allows to find the research gaps in the existing template security algorithms and suggests further development in the field of biometric template protection.

The emergence of deep learning in the field of computer vision has led to extensive deployment of convolutional neural networks (CNNs) in visual recognition systems for feature extraction. CNNs provide learning through hierarchical inferencing by providing multilayer architecture. Due to high processing capability of CNNs in multidimensional signals like images, they are considered to be predominant artificial neural networks. CNNs are extensively used in computer vision such as in image recognition where the intent is to automatically learn features followed by generalization and eventually recognizing the learned features. In this paper, we investigate the efficiency of CNNs: AlexNet and GoogLeNet under the effect of blurring which occurs frequently during image capturing process. Here, Gaussian blurring is used since it minimizes the noise embedded into the image. For experimentation, UC Merced Land Use aerial dataset is used to evaluate CNNs’ performance. The focus is to train these CNNs and classifying an extensive range of classes accurately under the influence of Gaussian blurring. Accuracy and loss are the parameters of classification considered for evaluating the performance of CNNs. Experimental results validated the susceptibility of CNNs towards blurring effect with GoogLeNet being more fluctuating to varied degrees of Gaussian blurring than AlexNet.

Today, world is paying more attention on those types of energy sources that create minimum pollution and fulfill the gap between demand and supply. Continuous increase in the power demand of consumers from various fields such as residential, industrial, and commercial, it is difficult to procure the additional supply from conventional sources with maintaining the pollution standard. So power-producing companies/agencies invest lot of fund in the development of such sources of energy which is nearly pollution-free and available from nature. In this regards, the alternative of conventional sources may be renewable energy sources. Out of different available renewable sources such as solar, wind, biomass, and small hydro, wind can be considered one of the good sources for the generation of power. Today, living standard of any country can be recognized by per capita energy consumption by its people. As power production from renewable sources may lead to minimum possible pollution, operating cost, and mostly freely and abundant availability, it will work as a major driving factor for some countries of the world to spend maximum available energy fund in the development of such mechanism/technique that will able to generate energy from renewable sources. Although power generation from wind has many advantages, major drawbacks are its intermittent nature, frequency instability, and continuous availability with certain threshold speed that is capable for power generation at all places. This paper describes the combined technique of PSO and ANN for forecasting of speed and power of wind to penetrate it in grid. The proposed method is applied on Indian wind power sector, and its results are compared with simple ANN and ANN-SVM methods.

As of late, the need for biometric security framework is elevated for giving safety and security against frauds, theft, and so on. Face recognition has gained a significant position among all biometric-based systems. It can be used for authentication and surveillance to prove the identity of a person and detect individuals, respectively. In this paper, a point-by-point outline of some imperative existing strategies which are accustomed to managing the issues of face recognition has been introduced along with their face recognition accuracy and the factors responsible to degrade the performance of the study. In the first section of this paper, different factors that degrade the facial recognition accuracy have been investigated like aging, pose variation, partial occlusion, illumination, facial expressions, and so on. While in the second section, different techniques have been discussed that worked to mitigate the effect of discussed factors.

Eyes are vital part of the body which can be affected by many diseases that lead to vision loss. Glaucoma is one such eye disease that may cause vision loss. There are multiple reasons for vision loss which may be due to the appearance of unwanted blood vessels that can be caused by high level of glucose in the blood composition. This abnormal growth or change in behavior of the blood vessels represents underlying indicators of problems associated with eye diseases such as diabetic retinopathy. Hence, early detection of eye ailments can be expedited with the help of various image processing technologies. The first step after image acquisition is the processing of images to extract features that exactly match the disease under observation. This paper attempts to evaluate the blood vessels using different segmentation algorithms and introduce an improved version of the vessel algorithm. The evaluation of segmentation approaches shows that Otsu clustering algorithm is performing best as compared to other state-of-the-art techniques using eye fundus images.

Multi-focus Image Fusion (MFIF) is a technique that combines multiple images to obtain a composite image in which all the objects are in-focus and have improved image quality. More information is stored by the focused image than that of the information stored by the source image. MFIF provides fused images which can be used for various image processing tasks like target recognition, feature extraction, and segmentation. There exists number of MFIF techniques in spatial as well as transform domain such as Stationary Wavelet Transform, Discrete Wavelet Transform, and Principal Component Analysis. In this paper, comparative analysis of various MFIF techniques which are used to fuse multi-focused images is done. Qualitative as well as quantitative evaluation has been carried out for various MFIF techniques. MFIF provides a fused image which helps for high resolution of vision. Various challenges/issues related to the existing MFIF techniques are also highlighted and will be helpful in the future.

Thyroid cancer is the common cancer which can be found mostly in women as compared to men around the world. Thyroid gland is a butterfly-shaped gland that is located around the voice box. Earlier, doctors used to evaluate thyroid cancer manually, but now they are using computer-aided diagnosis (CAD) system for automatic detection. As incidence rate of thyroid cancer is increasing day by day, therefore, a better technology is required for its earlier detection. There are different types of imaging modalities, such as magnetic resonance imaging (MRI), ultrasound (US), and computerized tomography (CT), which are utilized for early detection of diseases. This paper presents and discusses the major trends for an exhaustive overview of thyroid nodule detection, segmentation, classification, and feature extraction techniques. The approaches used in CAD are summarized with their advantages and disadvantages.

For computer vision, image matching is an essential trait which includes scene or object recognition. Detection using point feature method is much effective technique to detect a specific target instead of other objects or within clutter scene in an image. It is done by comparing correspondence points and analyzing between cluttered scene image and a target object in image. This paper presents novel SURF algorithm that is used for extracting, describing, and matching objects in colored images. The algorithm works on finding correspondence points between a target and reference images and detecting a particular object. Speeded-up robust features (SURF) algorithm is used in this study which can detect objects for unique feature matches and which has non-repeating patterns. This approach of detection can robustly find specified objects between colored cluttered images and provide constriction to other achieving near real-time performance.

This paper presents a smart and safe bike riding system to provide a safe and an intelligent driving features with accidental, speeding and rash driving alerts using fog computing. The system is based on the Ethernet-based 2nd Generation Intel Galileo Board. This intelligent system will be embedded in the upcoming bikes and motorcycles to prevent speeding, determine driver behavior and rash driving accidents. The whole idea of the system is to generate an alert to the user and provide caution alert to the user about their driving statistics and warn them as necessary. The system is embedded with various sensors like accelerometers, gyroscope, and GPS to make this system an intelligent one. The proposed outcome of the system aims as multiple benefits of preventing accidents, maintaining the ride statistics and getting the directions for the ride. Smart bike is an IoT-based ride system. In today’s world, everything is getting automated.

E-learning is one of the best tools to support an individual’s education system worldwide. There is a demand to create an e-learning site in Kannada for disabled people to remove accessibility barriers. In this paper, we have introduced a new framework for Web accessibility, providing easy access and lifelong learning to the Web site for Kannada’s deaf/blind speakers. The main purpose of this framework is to support Kannada’s deaf/blind by using the language of Kannada sign, speech text, and Kannada Moon code. Kannada-speaking disabilities can learn effectively in this framework.

An experimental study was conducted to predict the development and availability (DA) of the latest information and communication technology (ICT) and mobile technology (MT) in Indian and Hungarian University. A primary dataset with 328 instances and 16 features was analyzed using four supervised machine learning algorithms such as support vector machine (SVM), artificial neural network (ANN), random forest (RF), and logistic regression (LR). The dataset was trained and tested using hold out and K-fold cross-validation methods with classifiers. Further, to compare the performance of classifiers, T-test at 0.5 significant level was also applied. Feature mapping was also achieved by applying principal component analysis (PCA) to enhance the prediction accuracy of classifiers. The findings of the study conclude the feature extraction using PCA enhanced the prediction accuracy of each classifier except SVM with tenfolds at 0.5 thresholds of variance. Also, it is revealed that within the real time of 1.4 s ANN attained stable and highest accuracy in the prediction of DA of ICT and MT in the University of both countries. T-test implies the significant difference between RF and others in prediction accuracy. Also, a significant difference is found in ANN and others considering the processor time to train model for real-time prediction.

Web browsing is the need of the hour present state of the art problem of presenting the specific result to the users. This paper works towards creating an efficient search engine that removes stop words, extract meaningful words and form clusters of highest frequency words and final stage present the result in terms of URLs. The proposed system is divided into three phases: In the first phase, pre-processing is performed by eliminating the stop words. The outcome of this phase is a reduced query. In the second phase, the extraction of meaningful words with a frequent word or similar word replacement is applied. In the last phase, meaningful URLs are fetched through location-sensitive searching and then presented within the same interface. The result is presented in term of URLs fetched and execution time it takes to fetch the results.

Data mining is the procedure or technique of drawing out the facts and patterns hidden in huge sum of data and converts it into a readable and understandable form. Data mining has four main modules like classification, association rule analysis, and clustering and sequence analysis. The classification is the major module and is used in many different areas for classification problems. Classification process gives a summary of data investigation which may be utilized to develop models or structures, telling different classes or predict future data trends for improved understanding of the data at maximum. In this survey, various data mining classification techniques and some important data mining tools along with their advantages and disadvantages are presented. Data classification techniques are classified into three categories namely, Eager learners, Lazy learners, and other Classification techniques. Decision tree, Bayesian classification, Rule based classification, Support Vector Machines (SVM), Association rule mining and backpropagation (Neural Networks) are eager learners. The K-Nearest Neighbor (KNN) classification and Case Based Reasoning (CRT) are lazy learners. Other classification techniques include genetic algorithms, fuzzy logic and Rough Set Approach. Here six important data mining tools, basic Eager learner, Lazy learner and other classification techniques for data classification are discussed. The aim of this article is to provide a survey of six famous data mining tools and famous different data mining classification techniques.

In many real-world applications, class imbalance problem is the most attentive (also a major challenging) problem for machine learning (ML). The traditional classification algorithms assume evenly distributed in the underlying training set. In class imbalanced classification, the training set for one class called (majority class) far exceed the training set of the other class called (minority class), in which, the more often interesting class is minority class. We need to increase minority class samples than majority class samples in analysing of a datasets (related to an application). This is a hot problem in the past several decades. This article tries to provide as much as information to know or work about class imbalance problem with a detail description (from a user’s perspective). For this, we include several articles from a reputed publication like IEEE, ACM, Elsevier, Wiley, etc. Hence, this work will help a lot to all the future researchers to find out or a summary (about their interest) with respect to this class imbalance problem (raising in several applications).

Social media is a way of communicating with others and its popularity is growing worldwide. It has a lot of influence on its users. People read various posts and get affected by it. Suicide is one of the major health issues on social media which influence others to do the same. The number of suicides is increasing day by day. Thus, a need arises to find or develop a way to control suicides through social media. Machine learning is being widely used by many researchers for this purpose, with the help of psychiatrists. A lot of studies have been done in this field. In this paper, we have reviewed the existing work in this field inferring their limitations so that further work can be carried out.

Twitter is one of the fastest growing microblogging and online social networking site that enables users to send and receive messages in the form of tweets. Twitter is the trend of today for news analysis and discussions. That is why Twitter has become the main target of attackers and cybercriminals. These attackers not only hamper the security of Twitter but also destroy the whole trust people have on it. Hence, making Twitter platform impure by misusing it. Misuse can be in the form of hurtful gossips, cyberbullying, cyber harassment, spams, pornographic content, identity theft, common Web attacks like phishing and malware downloading, etc. Twitter world is growing fast and hence prone to spams. So, there is a need for spam detection on Twitter. Spam detection using supervised algorithms is wholly and solely based on the labelled dataset of Twitter. To label the datasets manually is costly, time-consuming and a challenging task. Also, these old labelled datasets are nowadays not available because of Twitter data publishing policies. So, there is a need to design an approach to label the tweets as spam and non-spam in order to overcome the effect of spam drift. In this paper, we downloaded the recent dataset of Twitter and prepared an unlabelled dataset of tweets from it. Later on, we applied the cluster-then-label approach to label the tweets as spam and non-spam. This labelled dataset can then be used for spam detection in Twitter and categorization of different types of spams.

The processing of structured data in Hadoop is achieved by Hive, a data warehouse tool. It is present on top of Hadoop and helps to analyze, query, and review the Big Data. The execution time of the queries has drastically reduced by using Hadoop MapReduce. This paper presents the detailed comparison of various optimizing techniques for data models like partitioning and bucket methods to improve the processing time for Hive queries. The implementation is done on data from New York Police Portal using AWS services for storage. Hive tool in Hadoop ecosystem is used for querying data. Use of partitioning has shown remarkable improvement in terms of execution time.

There has been much excitement recently about Big Data and the dire need for data scientists who possess the ability to extract meaning from it. Data scientists, meanwhile, have been doing science with voluminous data for years, without needing to brag about how big it is. But, now those large, complex datasets should process smartly. As a result, it improves productivity by reducing the computational process. As a result, Big Data analytics takes a vital role in intrusion detection. It provides tools to support structured, unstructured, and semi-structured data for analytics. Also, it offers scalable machine learning algorithms for fast processing of data using machine learning approach. It also provides tools to visualize a large amount of data in a practical way that motivates us to implement our model using scalable machine learning approach. In this work, we describe a scalable machine learning algorithm for threat classification. The algorithm has been designed to work even with a relatively small training set and support to classify a large volume of testing data. Different machine learning approaches implemented and evaluated using intrusion dataset. The data is normalized using the min–max normalization technique, and for SVM classification, data transforms into sparse representation for reducing computational time. Then using Apache Hive, we store the processed data into HDFS format. All the methods except the neural network are implemented using Apache Spark. Out of all the approaches, the fine KNN approach outperforms in terms of accuracy in a reasonable computational time, whereas the Bagged Tree approach achieves slightly less accuracy but takes less computational time for classifying the data.

In current times, the telecommunications industries are one of the tremendous factors to increase the economy, more so on the developing country like India. The telecom industry is one of the interesting industries to study, not only due to its technological development and its policies but also due to the high rate of development of this industry over the past few years and a significant factor to increase economy of the nation. Customer loyalty has become an important factor for both manufacturers and service providers in increasing competition for customers in today’s customer-centred era. The present study aims to understand the factors affecting customer loyalty. The findings indicated that information security, customer support, and responsiveness have a positive relationship between customer loyalty. Data are collected from 4 lakh tweets from Twitter by using popular hashtags and @ mention of telecommunication firms in Twitter. Topic modelling and sentiment mining were done on these tweets. The statistical analysis indicated that responsiveness, information security, and customer support play a significant factor for customer loyalty in the telecommunication industry.

The collaborative filtering is a very popular and powerful approach prominently being used in many research areas of computer science like recommendation systems, information retrieval, data mining, etc. When used for making recommendations, the traditional collaborative filtering methods suffer from certain problems, where the data sparsity is the significant one that causes the deterioration of the recommendation quality. In order to alleviate this issue, the research fraternity has started proposing the use of some additional domain information in formulating recommendations. In literature, different models have been proposed that make use of such kind of add-on information extensively and have also shown the promising performance than the other state-of-the-art approaches. Hence, the increasing use of add-on information is creating an overwhelming impact on the recommendation field. The piety of this article is to present a meticulous comparative study of various such recommendation models especially those which belong to the family of collaborative topic regression recommendation models in the light of several parameters and this study further leads to propose a novel recommendation prototype based on the fusion of different kinds of auxiliary domain knowledge.

With the growth of e-commerce, shopping online has now become a part and parcel of every one’s life. The advantage of e-commerce Web sites is that they can reach to a very large number of customers despite of distance and time limitations. The main aim of this paper is to extract the product information from various e-commerce sites. Extraction of such information can help the business organizations to fetch and attract the large number of customers to their Web site and increase profit. So, in this paper, we propose a fully automatic method which will extract and integrate information from multiple e-commerce Web sites in order to improve business decision making. The proposed method is also comparatively better at precision and recall than other methods.

Iptables is a stateful packet filtering firewall in Linux that monitors ingress and outgress traffic. The filtering is performed based on rules which are conditions predetermined by the network administrators. This paper investigates the performance of Iptables with different rule sizes (200, 500, 1000, 5000 and 10,000) and high traffic rates for different time durations. An experimental set-up is established for evaluating the performance of Iptables under stress by varying the packet rates from, viz., 1000 to 8000 PPS and different time durations (30–120 s). The performance is recorded on key parameters: CPU utilisation, response rate, packet dropped, packet processing time, throughput and bandwidth. These parameters reflect the sensitivity of the firewall for managing high rates of network traffic. ClassBench is used to generate rule sets of different sizes that imitate the real-life rule sets, and the network traffic is generated by DITG, a traffic-generating tool. Finally, a mathematical model is developed that can estimate the performance of the firewall in different traffic scenarios. Also, the proposed model is tested by performing validation tests on real test bed and shows less than 10% relative error.

An iris recognition framework is a standout, stable, secured, and widely accepted biometric in the secured authentication infrastructure. It has been broadly perceived as one of the grounded biometrics credited to its high precision execution. In this paper, a comprehensive analysis of several iris template security techniques has been carried out together with open opportunities for further research challenges. The analysis mainly pointed out that majority of the techniques results in a trade-off among various performance and security parameters, e.g., FAR, FRR, EER, revocability, diversity, and security. The well-known open iris databases such as CASIA, IITD were accessible to complete the experiments. It also has been noticed that the template security scheme for uni-biometric system may not work for multi-biometric system. Among the two broadly available template protection techniques, feature transformation schemes outperform its counterpart cryptosystems mainly due its unlinkability and irreversibility properties.

As Code Clones are defined on the notion of similarity in code fragments, it is necessary to first know what a code is meant by in accordance with Code Clones. A Source Code Fragment, which is a sequence of source code lines, is the basic entity that is used to analyze similarity/relation between Code Clones. For analysis, removal, avoidance, and management of Code Clones we have to first detect clones in software systems. There are more than 40 clone detection tools that implement some clone detection techniques to detect clones, but it is not well-defined what could be the appropriate minimum threshold for Clone length and with which unit of estimation. This paper, on the basis of Code Clone literature, presents different Units of Measurement of Clone Size and a comprehensive review of minimum Clone Size based on a particular technique used in Clone Detection and also argues that a unique Unit of Measurement and Minimum Clone Size should be presented.

Mobile edge computing (MEC) enables cloud-based services to extend to edge networks consisting of mobile base systems. MEC provides software and hardware platforms to incorporate seamless and decentralized data management schemes adjacent to base systems, thus reducing the end-to-end latency of the user. It is an integral component of the fifth-generation (5G) architecture and operates by providing innovative IT-based services. MEC spans across multiple authoritative domains where trust and interoperability among nodes is a prime concern between low power-enabled sensor nodes, as in the case of Internet of things (IoT)-based environments. The requirements of trust and interoperability make a blockchain framework applicable to MEC platform. In such platforms, miners can solve computationally expensive proof-of-work (PoW) puzzles containing mobile transactions as blocks added to immutable ledger so that a substantial amount of CPU computations and energy constraints are consumed. This article presents a systematic survey of MEC architecture and introduces a mobile blockchain framework that can be incorporated with the MEC architecture to facilitate the mining scheme. Then, the article analyzes the effects of integration of blockchain with MEC platform. Finally, concluding remarks and future work are provided.

Network intrusion detection systems (NIDS) are emerging as a reliable solution in providing protection against threats to integrity and confidentiality of the information on the Internet. Two widely used open-source intrusion detection systems are Snort and Suricata. In this paper, Snort and Suricata are compared experimentally through a series of tests to identify more scalable and reliable IDS by putting the systems under high traffic. Results indicated that Snort had a lower system overhead than Suricata and utilized only one processor on a multi-core environment. However, Suricata evenly utilized all the processing elements of the multi-core environment and provided higher packet analysis rate. For malicious traffic, both Snort and Suricata dropped packets with Snort on the higher side for low traffic rate and size. But with large packet size and high rate of malicious input traffic, Suricata dropped more packets as compared to Snort. It was also observed that the memory utilization of Suricata depended on both the size of traffic and the amount of malicious traffic; whereas, memory utilization of Snort was independent of the input traffic.

Today, human still carry number of cards to authenticate their identities across the globe. Smart cards possess potential to substitute all other existing IDs by smart card IDs at national and global level. Contemporary issues of terrorism and illegal migrants across many international borders too support the cause of seeking a viable solution. This paper has examined the evolution and necessity of human identification along with current worldwide scenario of existing national IDs. Integration of biometrics with smart card technology presents a strong authentication tool to the identity card holder. This paper offers an insight into the feasibility and technological aspects of this potential application. A recent trend of using radio-frequency identification (RFID) and biometric technologies for personal identification in e-passports and other applications too paves the way to explore a global identity solution [1]. However, apart from the technological challenges, there are policy and legal constraints imposed by various governments across the globe which acts as a barrier and has been briefly touched upon in this paper. In order to accommodate the massive global IDs, an IPv6-based numbering scheme has been proposed for identity registration and data access of every human across the globe in this research paper. The state-of-art technologies and IoT with its widespread usage facilitates this proposal of integrated global ID solution incorporating smart card with RFID and biometric technologies against a multipurpose universal ID framework [2]. This study has also highlighted the future promises and the major research impact of the proposed application.

With the advancement in the VLSI technology, the demand for low power consumption and high performance increased gradually. When the applications of data retention are considered, then the need for advanced memory units is taken into account. This requirement of enhanced memory units is incorporated with the concept of conservation of energy which is achieved by using low-power techniques. In digital circuits, flip-flops are the essential memory and timing elements. New methods and techniques needed to be developed for implementing energy-efficient low-power flip flops. This paper proposes dual edge-triggered flip-flop (DETFF) along with gating technique, one of the most reliable low-power techniques which provide one-time solution to low-power applications. The DETFF circuit based on gating technique is simulated using MENTOR GRAPHICS tool in 180 nm technology. This design is efficient in reducing power dissipation leading to the reduction in area and delay and subsequently leads to the high speed of the device.

Slot games are games consisting of reels, appearing on the display randomly after the triggering of the spin symbol or the lever. It has various types of reel sets, most commonly used being 5 * 3. The game was played by inserting a coin, a ticket or a barcode in the slot provided in the machine. Thereafter, the reels displayed on the screen would start spinning for some time and then stop. As the reels will stop the symbols that will then be displayed on the screen, may–may not forms a winning pattern or a sequence. The player can either lose the bet that he played or win higher bonus depending on the sequence that gets displayed on the screen after the reels spin. Therefore, the basic principle behind all the casino games is a chance. This work throws light upon the development of the slot game which is coded in Java using various advanced Java techniques and Web server protocols.

India, the land of farmers, where agriculture has always been the primary occupation of the people, more than 50% of the population is still engaged in agriculture and its allied sectors. However, over the years, a significant rapid decline has been observed in the contribution by the agriculture sector toward India’s GDP rate. In this paper, we aim toward identifying the gap between the ratio of high inputs and low yields by portraying the various loopholes in traditional Indian agriculture methods and how hydroponic agriculture is need of the hour for the growth of Indian agriculture. Also, the current trends in technology and research in the field of hydroponics around the world have been discussed to show how it can provide an ideal solution to the insufficiency of traditional farming, and how Indian farmers can adopt its implementation practices to boost their crop yield and income. Also, an IoT-based application has been proposed for monitoring and control of a hydroponic setup.

In India, presently the meritorious sportspersons are not being benefited due to the non-implementation of the latest technologies; this can be achievable with the sports policy using IoT platform. By developing a special web address or a mobile application for each game to have the live telecast matches, an online score of matches with player details and along with online referee support method interlinking these through the internet and store all the information in the cloud accessible only to the sports policy authorities. Within this context, the contribution of this study is (i) Educating player in problem-solving for better performance. (ii) Player healthcare like nutrition, medical supports, fitness. (iii) High quality of digital and visualization technological skillful game training. (iv) Latest referee rules and regulations. (v) Identification of the meritorious sportsperson from the metropolitan area to agency area. This all is monitored to provide sports Excellency facilities and required sports benefits.

Electing a Prime Minister (PM) is a process that occurs every 5 years in India, through conducting an election. Who will win the election is one of the most asked questions on everyone’s tongue, as the election is just few months away. The objective of this research is to predict the likelihood of PM Narendra Modi’s chances to continue as the Prime Minister of India using Bayesian network approach. The aim of this research is not to develop a new predictive algorithm, but to use the existing approach for making predictions on a real-life scenario. Our Bayesian modeling is based on public responses available on social media, India statistics, and news articles on the key policies undertaken by PM Narendra Modi during his current tenure. We explore using causal and diagnostic reasoning to find new insights on the factors shaping his win or no-win, verdict on his government strength and weakness. Our Bayesian model reveals that the current Prime Minister Modi has 61.4% chances of winning the upcoming 2019 elections.

In recent years, there has been a lot of active research on energy recovery from waste heat obtained from various sources. A remarkable potential for harvesting the energy lies in the waste heat obtained during households daily chores. In this paper, waste heat obtained from chulhas has been exploited to generate energy with the help of thermoelectric energy system. Thermoelectric generators convert the waste heat to voltage are coupled with suitable heat sink, interfaced with DC–DC converter and storage circuit. The thermoelectric systems for waste heat recovery have low system efficiency, which prevents using them feasibly for the direct fuel energy conversion. This paper focuses on design of various key parameters, components and factors that determine the performance of DC–DC converter/energy harvester system. The theoretical results obtained showed good agreement compared with the simulation results using LTSPICE. This is a step toward greener source of energy as it does not hinder the normal working of the chulhas and the waste heat which is usually emitted out in the atmosphere and is of no use.

Innovations in pedagogy and technology have lead to a new paradigm in teaching particularly in higher education. At the nexus of this new paradigm is blended learning and e-learning. Blended learning refers to combination of synchronous and asynchronous learning activities. While e-learning refers to learning system that utilize electronic means or information communication technology (ICT) to deliver information for education or training purposes. Though in infancy e-learning in India has garnered pace in last decade, introduction of SWAYAM portal for offering massive open online courses (MOOCs) has been an important initiative by Government of India in this direction. The present study conceptualizes various factors of e-Learner Satisfaction into one model. The data for study was collected from students of north Indian universities. A total of 266 responses were recorded out of which 25 responses were eliminated due to incomplete information. Final data analysis was conducted using 241 responses using structural equation modeling (SEM) (Amos 20). From the data, it was observed the constructs (predictors) namely Instructors Attitude, Learners Attitude, Course Quality, Technology Quality, Assessment Quality and Perceived ease of use explained 66.2% variance (adjusted R2 = 66.2% and p < 0.05) in Perceived e-Learner Satisfaction.