Vellore Institute of Technology (VIT) is a private university located in&nbsp;Tamil Nadu, India. Founded in 1984, as Vellore Engineering College, the institution offers 20 undergraduate, 34 postgraduate, four integrated and four research programs. It has campuses in Vellore, Amravati, Bhopal and Chennai.

VIT is one of the top ranked private universities in India according to NIRF, THE and QS Rankings.&nbsp;Govt. of India has recognized&nbsp;VIT, Vellore as an&nbsp;Institution of Eminence. This has allowed VIT to take independent quality initiatives and move up in world ranking.

&nbsp;

&nbsp;

VIT University

Department of Analytics

Location estimation is one of the critical requirement for developing smart environment products. Due to huge utilization and accessibility of WiFi infrastructure facility in indoor environments, researchers widely studied this technology to locate users accurately to provide several services instantly. In this research work, a hybrid algorithm namely fuzzy decision tree (FDT) with evolutionary fuzzy clustering methods is adopted for optimal user localization in a closed environment. Here we consider the wireless signal strengths received from the smart phones as predictors and the location of the user as the classification label. The required data for the current research is collected from the physical facility available at an office location in USA. The classification results obtained are promising enough to show that the evolutionary clustering approaches provide good fuzzy clusters for FDT induction with better accuracy. © 2021 Wiley Periodicals LLC

Department of Software Systems

School of Computer Science and Engineering

Vellore Campus

World University root node

Department of Information Security

International Journal of Intelligent Systems

Fuzzy decision trees embedded with evolutionary fuzzy clustering for locating users using wireless signal strength in an indoor environment

This paper focuses to solve two important problems faced in this digital era. The first problem is handling large amount of digital data. Digital sectors are finding difficulty in handling repetitive works and standardization of process. Our University is a eco-friendly, digital university each and every work of our students are maintained digitally. So for a course instructor to open, view, correct and finding plagiarism in the assignment or homework submitted by students is a tough task. So in this paper this tough task is simplified by applying Robotic Process Automation (RPA). The process of downloading and reading the students’ digital document is automated. The second problem is the plagiarism. There has been a massive threat in the academic and technological integrity of innovation in the name of paraphrasing. In a survey conducted by Donald McCabe of Rutgers University over 70 percent of students admit to paraphrasing and copying of others work. They have also fabricated bibliography. This problem can be addressed using the state-of-the-art document similarity deep learning neural networks. In this paper, we propose the design and implementation of a Siamese Neural Network (SNN), Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) to identify paraphrasing, using only features of similarity between words where the dataset is evolutionary and the model keeps learning with incremental sources of the document’s plagiarism. This paper will also provide the Tensorflow based implementation of such a model. The proposed system will predict the similarity between documents and predict the plagiarism percentage.

Department of Database Systems

Solid State Technology

Plagiarism Detection using Attention based Text Similarity

Nowadays, web applications have become most prevailing in the industry, and the critical data of most organizations are stored using web apps. Thus, web applications pose a large target for assorted cyber-attacks. As mitigation for them, among many proposed solutions, web application honeypots are much sophisticated and robust protection.In this paper, we propose a low interaction, adaptive, and dynamic web application honeypot that imitates the vulnerabilities through HTTP events. The honeypot is built with SNARE and TANNER; SNARE devises the attack surface and sends the requests to TANNER, which evaluates them and determines how SNARE should respond to the requests. TANNER is an analysis and classification tool, which analyses and evaluates HTTP requests served by SNARE, and composes a dynamic response by its emulation engine. The honeypot works abreast major complex vulnerabilities, deceiving the attacker by lying under the surface of a real web application.

International Journal of Innovative Technology and Exploring Engineering (IJITEE)

An Innovative Security Strategy using Reactive Web Application Honeypot

In recent years, the usage of wireless devices and wireless service has been increased exponentially and it results in spectrum scarcity. The policies of regulatory authority employ static spectrum allocation methods and assign new spectrum band for offering new kind of services to the users. These approaches lead to poor utilization of available spectrum bands. The cognitive radio (CR) provides better solution to these problems and it mainly focuses on efficient utilization of available spectrum bands. CR network (CRN) has to adopt spectrum management techniques to assign the unused spectrum band to the CR users by following a sequence of actions such as spectrum sensing, decision and management. Spectrum sensing is a vital process in spectrum allocation. In the traditional approaches, sensing accuracy is brought down by the probabilities of misdetection and false alarm rate and hence, sensing accuracy becomes low. As a result, the Cognitive Users (CUs) face the challenge of prolonged time to complete perfect cognitive radio communication. To overcome this issue, a cooperative spectrum sensing technique with a characteristic based cluster classifier has been proposed. This classifier learns the states and their transitions in the radio frequency environment, as well as the primary user activities at regular time intervals to support the spectrum decision technique. The novelty of the work is to propose a hybrid approach which combines clustering with expected maximization (EM) algorithm and reinforcement learning (RL) techniques. This hybrid approach enhances the system performance with accurate sensing results and by identifying the optimum spectrum band through hierarchical access model using interweaving approach, energy consumption is minimized. The simulation results show that by decreasing the probabilities of error ratio, false alarm rate and missed detection, the accuracy of sensing results is improved. Further, this hybrid approach outperforms the traditional approaches in terms of probability of detection even in low SNR values. © 2020 Elsevier B.V.

Computer Networks

A hybrid spectrum sensing approach to select suitable spectrum band for cognitive users

Cyberattacks are increasing rapidly with rapid Internet advancement and, the cybersecurity situation is not optimistic. Anomaly detection is one of the challenging sectors of network security, which shows a significant role in any organization. Many anomaly detection systems identify malicious activities by deploying machine learning and deep learning techniques. The major contribution of this research is to develop an anomaly detection model for networks using a homogenous ensemble of Long-Short-Term-Memory integrated with Genetic Algorithm (GA) utilized for feature extraction. An extensive literature on anomaly detection, which utilizes deep learning algorithms, is studied. NSL-KDD and UNSW-NB datasets are deployed for evaluating the proposed network anomaly model. The experimental analysis shows that the proposed ensemble is superior to other ensembles with a maximum accuracy of 99.9% and a minimum false alarm rate of 1.56% on NSL-KDD dataset and a maximum accuracy of 99.3% is obtained on UNSW-NB15 dataset with false alarm rate of 1.7%. Hence, the proposed model performs fair on both the datasets. © 2020 John Wiley &amp; Sons Ltd

Department of Software and Systems Engineering

School of Information Technology and Engineering

Transactions on Emerging Telecommunications Technologies

An intelligent ensemble of long-short-term memory with genetic algorithm for network anomaly identification

The incidence of skin cancer has been increasing in recent years and it can become dangerous if not detected early. Computer-aided diagnosis systems can help the dermatologists in assisting with skin cancer detection by examining the features more critically. In this article, a detailed review of preprocessing and segmentation methods is done on skin lesion images by investigating existing and prevalent segmentation methods for the diagnosis of skin cancer. The pre-processing stage is divided into two phases, in the first phase, a median filter is used to remove the artifact; and in the second phase, an improved K-means clustering with outlier removal (KMOR) algorithm is suggested. The proposed method was tested in a publicly available Danderm database. The improved cluster-based algorithm gives an accuracy of 92.8% with a sensitivity of 93% and specificity of 90% with an AUC value of 0.90435. From the experimental results, it is evident that the clustering algorithm has performed well in detecting the border of the lesion and is suitable for pre-processing dermoscopic images. Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

International Journal of e-Collaboration

Border detection in skin lesion images using an improved clustering algorithm

In order to cater to the energy requirement in the form of steam at a reasonable cost, the process industries are relying on the waste incineration plants by engaging themselves through industry symbiosis. However, before the establishment of industrial symbiosis, it is very crucial to monitor and predict the operational performance of the boiler used in the waste incineration plants. The existing works focus on using Artificial Neural Networks (ANNs) for prediction of the performance of the boiler in terms of pressure, temperature, and mass flow rate of steam using the input parameters viz. feed water temperature, feed water pressure, incinerator exit temperature and conveyor speed. However, the problem with this approach is that shallow ANNs cannot model the complex mathematical non-linear relationships so precisely. In addition, ANNs are not interpretable which makes stakeholders apprehensive to use these networks in production. In this paper, we address these drawbacks of ANNs by modeling the complex relationship governing the boiler performance by using a set of machine learning and deep learning models. Also, the research paper introduces multiple techniques like feature importance, Partial Dependence Plots(PDP) plots etc. which interpret the reason behind the model's output to make it more reliable for the stakeholders. It has been empirically shown that the new Machine Learning and Deep Learning models performed better than the ANNs for predicting the boiler performance. The Random Forest model made a Mean Absolute Percentage Error (MAPE) of 1.12 and LSTMs had a MAPE of 1.14 in the prediction of steam temperature C^o which is a significant improvement in comparison to the original ANN model which had a MAPE of 6.93. In the case of the predictions for steam pressure kgf/cm2 the MAPE for the Random Forest model and LSTM was 5.54 and 4.21 respectively as compared to ANNs MAPE of 1.49. Similarly for steam mass flow rate(t/h), the MAPE was improved to 15.6 and 9.63 by Random Forest Model and LSTM respectively, which was originally 18.77 for ANN based model. These results clearly show that LSTM based models outperformed ANNs and Random Forests in terms of prediction accuracy. © 2013 IEEE.

Fulltext

IEEE Access

Performance Prediction and Interpretation of a Refuse Plastic Fuel Fired Boiler

Age and gender prediction from facial images with high accuracy is of immense importance in various fields viz. social media, retail business, and academic research. In this paper, we make age prediction with an optimized model for efficient training by wide residual networks (ResNet) and efficient gradient optimization on loss function for better test accuracy. Our model has been evaluated and tested on IMDb-WIKI and APPA-REAL dataset and it performed well in evaluation compared to traditional deep Convolutional Neural Networks such as VGG-16, naive SVM classifiers with feature edge parameters, etc. The final accuracy achieved for our model is 96.269% with the wide ResNet architecture along with augmentation and erasing techniques on images. Significant reduction in the Mean Apparent Error (1.73) on apparent images and Mean Apparent Error (1.65) on Real images of the APPA-REAL dataset has been achieved with respect to traditional VGG-16 model. © 2020, Springer Nature Singapore Pte Ltd.