f) 2020 - 82 Documents
Permanent URI for this communityhttps://dspace.psgrkcw.com/handle/123456789/3917
Browse
10 results
Search Results
Item OVERLAPPING COMMUNITY STRUCTURE DETECTION USING TWITTER DATA(International Journal on Emerging Technologies, 2020) Sathiyakumari, K; Vijaya, M SOverlapping community detection is progressively becoming a significant issue in social network analysis (SNA). Faced with massive amounts of information while simultaneously restricted by hardware specifications and computation time limits, it is difficult for clustering analysis to reflect the latest developments or changes in complex networks. Techniques for finding community clusters mostly depend on models that impose strong explicit and/or implicit priori assumptions. As a consequence, the clustering effects tend to be unnatural and stay away from the intrinsic grouping natures of community groups. In this method, a process of enumerating highly cohesive maximal community cliques is performed in a random graph, where strongly adjacent cliques are mingled to form naturally overlapping clusters. These approaches can be considered as a generalization of edge percolation with great potential as a community finding method in real-world graphs. The main objective of this work is to find overlapping communities based on the Clique percolation method. Variants of clique percolation method such as Optimized Clique percolation method, Parallel Clique percolation method have also been implemented. This research work focuses on the Clique Percolation algorithm for deriving community from a sports person’s networks. Three algorithms have been applied for finding overlapping communities in the sports person network in which CPM algorithm discovered more number of communities than OCPM and PCPM. CPM overlapping algorithm discovered 198 communities in the network. OCPM algorithm found 180 different sizes of communities. PCPM algorithm discovered 170 communities and different size of the node in the graph. The community measures such as size of the community, length of community and modularity of the community are used for evaluating the communities. The proposed parallel method found a large number of communities and modularity score with less computational time. Finally, the parallel method shows the best performance is detecting overlapping communities from the sports person network.Item GATED RECURRENT NEURAL NETWORK FOR AUTISM SPECTRUM DISORDER GENE PREDICTION(International Journal on Emerging Technologies, 2020) Pream Sudha, V; Vijaya, M SAutism Spectrum Disorder (ASD) is the fastest-growing complex disorder and the genetic ground of this comprehensive developmental disability is very difficult to research. Autism diagnosis for an average child is not done till the age of four, though it can be given at the age of 18 months to two years. Hence a computational model that enables the early diagnosis and personalized treatment is the need of the hour. In this research work, a deep learning based approach is proposed for Autism Spectrum Disorder (ASD) gene prediction. There are various contributors for Autism including genes, mutations, chromosomal settings influence of the environment, prenatal influences, family factors and problems during birth. Recurrent Neural Network (RNN) based Gated Recurrent Units (GRU) are implemented to develop a model that predicts ASD genes, mutations and gene susceptibility. GRUs with their internal memory capability are valuable to store and filter information using the update and reset gates. Also GRU offers a powerful tool to handle sequence data. The model is trained using three simulated datasets with features representing genes, mutations and gene susceptibility to ASD. Besides, the proposed approach is compared to original RNN and Long Short Term Memory Units (LSTM) for ASD prediction. The experimental results confirm that the proposed approach is promising with 82.5% accuracy and hence GRU RNN is found to be effective for ASD gene prediction.Item BUILDING ACOUSTIC MODEL FOR PHONEME RECOGNITION USING PSO-DBN(Inderscience Enterprises Ltd, 2020-04-02) Laxmi Sree, B R; Vijaya, M SDeep neural networks has shown its power in generous classification problems including speech recognition. This paper proposes to enhance the power of deep belief network (DBN) further by pre-training the neural network using particle swarm optimisation (PSO). The objective of this work is to build an efficient acoustic model with deep belief networks for phoneme recognition with much better computational complexity. The result of using PSO for pre-training the network drastically reduces the training time of DBN and also decreases the phoneme error rate (PER) of the acoustic model built to classify the phonemes. Three variations of PSO namely, the basic PSO, second generation PSO (SGPSO) and the new model PSO (NMPSO) are applied in pre-training the DBN to analyse their performance on phoneme classification. It is observed that the basic PSO is performing comparably better to other PSOs considered in this work, most of the time.Item ASSOCIATION RULE MINING FOR CLIQUE PERCOLATION ON COMMUNITY DETECTION(International Journal of Advanced Science and Technology, 2020) Sathiyakumari, K; Vijaya, M SThe recognition of communities linking like nodes is a demanding subject in the revision of social network data. It has been extensively considered in the social networking community in the perspective of underlying graph structure besides communication among nodes to progress the eminence of the discovered communities. A new approach is proposed based on frequent patterns and the actions of users on networks for community detection. This research work spends association rule mining to discover communities of similar users based on their interests and activities. The Clique Percolation technique initially anticipated for directed networks for driving communities is enlarged by using the ascertained prototypes for seeking network components, i.e., internally tightly linked groups of nodes in directed networks discovering overlapping communities efficiently. The community measures such as the bulk of the community, piece of community and modularity of the community are used for testing the reality of communities. It tests the proposed community detection approach using a sample twitter data of sports person networks with F-measure and precision showing that the proposed method principals to improve the community detection quality.Item DEEP LEARNING PREDICTIVE MODEL FOR DETECTING HUMAN INFLUENZA VIRUS THROUGH BIOLOGICAL SEQUENCES(Springer Link, 2020-09-08) Nandhini, M; Vijaya, M SSwine influenza is a contagious disease which is generated by one of the swine influenza viruses. Any modification in protein will alter the biological activity and lead to illness. Obtaining appropriate information from virus protein sequence is an interesting research problem in bioinformatics. The aim of this research work is to develop deep neural network (DNN)-based virus identification model for detecting the virus accurately with the protein sequences using deep learning. Deep learning is gaining more importance because of its governance in terms of accuracy when the network trained with large amount of data. A corpus of 404 protein sequences associated with nine types of human influenza virus is collected for training the deep neural network and building the model. Various parameters of the DNN such as input layer, hidden layer and output layer are fine-tuned to improve the efficiency of the model. Sequential model is created for developing DNN classification model using Adam optimizer with Softmax and ReLu activation functions. It is observed that experiments of proposed human influenza virus identification model with DNN classifier give 80% of accuracy and outperform with other ensemble learning algorithms.Item DEEP NEURAL NETWORK FOR EVALUATING WEB CONTENT CREDIBILITY USING KERAS SEQUENTIAL MODEL(Springer Link, 2020-08-08) Manjula, R; Vijaya, M SWeb content credibility determines the measure of acceptable and reliable of the web content that is observed. Content will prove to be unreliable if it is not updated, and it is not controlled for remarkable, and therefore, web content credibility is considerably essential for the people to assess the content. The analysis of content credibility is a vital and challenging task as the content credibility is outlined on crucial factors. This paper focuses on building deep neural network (DNN)-based predictive model using sequential model API to evaluate credibility of a webpage content. Deep neural network (DNN) is considered as an extremely promising decision-making architecture, and it performs feature extraction and transformation with the use of refined statistical modeling. A corpus of 400 webpage contents has been developed, and the factors like readability, freshness, and duplicate content are defined and captured from the webpage content. These features are redefined, and a new set of features is self-learned through the deep layers of neural network. Numeric labeling is used for defining credibility, wherein five-point Likert scale rating is used to denote the content credibility. By using sequential model, KerasRegressor with ADAM optimizer and a multilayer network is generated for building DNN-based predictive model and discovered that deep neural network outperforms other general regression algorithms in prediction scores.Item MALWARE FAMILY CLASSIFICATION MODEL USING USER DEFINED FEATURES AND REPRESENTATION LEARNING(Springer Link, 2020-11-20) Gayathri, T; Vijaya, M SMalware is very dangerous for system and network user. Malware identification is essential tasks in effective detecting and preventing the computer system from being infected, protecting it from potential information loss and system compromise. Commonly, there are 25 malware families exists. Traditional malware detection and anti-virus systems fail to classify the new variants of unknown malware into their corresponding families. With development of malicious code engineering, it is possible to understand the malware variants and their features for new malware samples which carry variability and polymorphism. The detection methods can hardly detect such variants but it is significant in the cyber security field to analyze and detect large-scale malware samples more efficiently. Hence it is proposed to develop an accurate malware family classification model contemporary deep learning technique. In this paper, malware family recognition is formulated as multi classification task and appropriate solution is obtained using representation learning based on binary array of malware executable files. Six families of malware have been considered here for building the models. The feature dataset with 690 instances is applied to deep neural network to build the classifier. The experimental results, based on a dataset of 6 classes of malware families and 690 malware files trained model provides an accuracy of over 86.8% in discriminating from malware families. The techniques provide better results for classifying malware into families.Item MALWARE DETECTION USING GIST FEATURES AND DEEP NEURAL NETWORK(IEEE Xplore, 2020-04-23) Krithika, V; Vijaya, M SMalware is a virus file which causes damage to system files like executable files, documents, program files. This intent affects the performance of the system. Malware detection is vital with occurrence of malicious code on the internet and it provides an early warning for the computer security regarding malware and cyber-attacks. Real time malware detection is still a challenge though there is a considerable research showing advances in methods that can automatically predict the malicious of a specific file, program. Though the existing malware scanner can recognize the infected file, it produces the conflicting decisions and the accuracy of prediction is still not promising. Hence it is proposed to develop an accurate malware identification model using intelligent learning method. In this paper, malware detection problem is formulated as binary classification task and appropriate solution is obtained using machine learning. A database consisting of 400 executable files of which 200 virus samples and 200 benign samples have been used to prepare the training dataset. All the executable files have been converted into gray scale images from which the GIST features are derived. The contemporary deep learning is adopted to build the binary classifier which takes the GIST features as input. The experimental results provide an accuracy of over 81% in discriminating malware and benign files. It is reported that deep neural network based binary classification achieved improved predictive performance when compared with supervised learning.Item MEASURING WEB CONTENT CREDIBILITY USING PREDICTIVE MODELS(Springer Link, 2020-01-30) Manjula, R; Vijaya, M SWeb content credibility is a measure of believable and trustworthy of the web content that is perceived. Content can turn out to be unreliable if it is not up-to-date and it is not measured for quality or accuracy and therefore, web content credibility is important for the individuals to access the content or information. The analysis of content credibility is an important and challenging task as the content credibility is expressed on essential factors. This paper focus on building predictive models to discover and evaluate credibility of a web page content through machine learning technique. A corpus of 300 web page contents have been developed and the factors like Readability, Freshness, Duplicate Content are defined and captured to model the credibility of web content. Two different labeling such as binary labeling and numeric labeling are used for defining credibility. In case of binary labeling, the high and low credibility of web content are represented by 1 and 0, respectively, whereas in case of numeric labeling five-point scale rating is used to mark the content credibility. Accordingly, two independent datasets have been developed. Different regression algorithms such as Linear Regression, Logistic Regression, Support Vector Regression (SVR) are employed for building the predictive models. Various experiments have been carried out using two different datasets and the performance analysis shows that the Logistic Regression model outperforms well when compared to other prediction algorithms.Item DISCOVERING HUMAN INFLUENZA VIRUS USING ENSEMBLE LEARNING(Elsevier, 2020) Nandhini, M; Vijaya, M SSwine influenza is an infectious disease caused by one of the swine influenza viruses. The swine flu is also seen in humans and it is caused by human influenza viruses. Extracting useful information from virus protein sequences is an interesting research problem. Any changes in protein will alter the biological function and cause disease. The main objective of the research is to build a classification model that will discover the human influenza virus using its protein sequences through ensemble learning classifiers. A total of 404 protein sequences related to various types of human influenza virus were selected for our study. The classification models were built using ensemble learning techniques, such as bagging, boosting, voting, and forest of randomized trees. The accuracy of the classifiers was evaluated and the results were reported. Boosting and randomized trees classifiers were effective in recognizing the human influenza virus.