Prashant Singh Rana

Self-Healing Neural Model for Stabilization Against Failures Over Networked UAVs

Unmanned aerial vehicles (UAVs) allow formation of wide range ad hoc networks. These ad hoc formations with unmanned vehicles provide coverage of vast areas of applications involving mission dependent activities. Such networks can solve various issues related to civilian and military activities. One of the main applications of these networks is continuous surveillance. Surveillance by multiple nodes in ad hoc mode is directly dependent upon the continuous data sharing, cooperative decision making and stabilized network formation. Failures in network can hinder the performance and can decrease its operability. It is difficult to aloof network from discrete failures. Therefore, stabilized model is required which can provide stability to the whole network. For this, a self-healing neural model is developed which is capable of handling uncertain failures. It also provides provision for recovery of nodes from failure to stabilized state.

Ensemble approach to detect profile injection attack in recommender system

Recommender systems apply knowledge discovery techniques to specific problem of making personalized recommendation for the products or services to the users. The huge growth in the information and the number of visitors to the web sites especially on e-commerce in last few years creates some challenges for recommender systems. E-commerce recommender systems are highly vulnerable to the profile injection attacks, involving insertion of fake profiles into the system to influence the recommendations made to the users. Prior work has shown that even a small number of malicious profiles can bias the system significantly. In this paper, we compare six machine learning algorithms and based on their performance we build our ensemble model and measure its performance in the detection of profile injection attacks.

B 2 FSE framework for high dimensional imbalanced data: A case study for drug toxicity prediction

Abstract The life of people is imperiled by umpteen chemicals unwittingly through the diverse sources like food, cleaning products, medicines, etc. At times, these chemicals can be toxic. Assessing and analyzing the toxicity of these chemical compounds can lead us to prospects to improve the environmental chemicals and invent new medicines. Tox21 crowdsourcing program initiative brings an evolutionary breakthrough for the researchers to develop better toxicity assessment techniques. Machine learning has received much attention in the domain of predictive analytics as it applies computational statistics and offers automation environment to expedite the data modeling process. The goal is to develop an efficient prediction model, combined with the machine learning algorithmic characteristics, which can predict whether a chemical compound is toxic and can affect the health adversely or not. Hence, an efficient pre-processing method should be adopted to achieve the best performance of the machine learning classifier. This work is a specific case study which proposes a Better Balanced Feature Selection Ensemble(B2FSE) framework for the classification of drug toxicity molecules, carried out on imbalanced and high dimensional complex drug data. We show that, an ensemble feature selection and an ensemble classifier, integrated with random subset selection, and a class balancer have the potential to generate more accurate, lower cost, and balanced classification framework. The performance of the proposed framework, when evaluated with different evaluation parameters and compared with the state-of-the art methods like SVM, random forest, bagging, etc., is found to be superior than the available methods.

Prediction of drug synergy score using ensemble based differential evolution

Prediction of drug synergy score is an ill-posed problem. It plays an efficient role in the medical field for inhibiting specific cancer agents. An efficient regression-based machine learning technique has an ability to minimise the drug synergy prediction errors. Therefore, in this study, an efficient machine learning technique for drug synergy prediction technique is designed by using ensemble based differential evolution (DE) for optimising the support vector machine (SVM). Because the tuning of the attributes of SVM kernel regulates the prediction precision. The ensemble based DE employs two trial vector generation techniques and two control attributes settings. The initial generation technique has the best solution and the other is without the best solution. The proposed and existing competitive machine learning techniques are applied to drug synergy data. The extensive analysis demonstrates that the proposed technique outperforms others in terms of accuracy, root mean square error and coefficient of correlation.

Fraudulent Firm Classification: A Case Study of an External Audit

ABSTRACT This paper is a case study of visiting an external audit company to explore the usefulness of machine learning algorithms for improving the quality of an audit work. Annual data of 777 firms from 14 different sectors are collected. The Particle Swarm Optimization (PSO) algorithm is used as a feature selection method. Ten different state-of-the-art classification models are compared in terms of their accuracy, error rate, sensitivity, specificity, F measures, Mathew’s Correlation Coefficient (MCC), Type-I error, Type-II error, and Area Under the Curve (AUC) using Multi-Criteria Decision-Making methods like Simple Additive Weighting (SAW) and Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). The results of Bayes Net and J48 demonstrate an accuracy of 93% for suspicious firm classification. With the appearance of tremendous growth of financial fraud cases, machine learning will play a big part in improving the quality of an audit field work in the future.

Information Retrieval in Web Crawling Using Population Based, and Local Search Based Meta-heuristics: A Review

The exponential growth and dynamic nature of the world wide web has created challenges for the traditional Information Retrieval (IR) methods. Both issues are the imperative source of problems for locating the information on web. The crawlers expedite web based information retrieval systems by following hyperlinks in web pages to automatically download new and updated content. The web crawlers systematically traverse the web pages, and fetch the information viz. nature of the web content, hyper-links present on the web page, etc. This paper reviews and compares the meta-heuristic approaches like population based, evolutionary algorithms, and local search used for IR in web crawling. This paper reviews how these techniques has been developed, enhanced and applied.

Parameter Optimization for H.265/HEVC Encoder Using NSGA II

High Efficiency Video Coding (H.265/HEVC) is the latest technology standard proposed by Joint Collaborative Team on Video Coding (JCT-VC). There are quite a few parameters for this encoder required to accomplish this goal. If a single standard configuration file is used for all genres of videos that may not maintain the optimal quality in all encoded videos. This is because every video has objects with unlike speeds of movement. Therefore, encoding factors must be customized in the most favorable way for each video separately. The work propose here is to use NSGA II for multi-objective optimization in order to find out the respective personalized encoding parameters to obtain higher Compression Ratio and Peak Signal-to-Noise Ratio (PSNR). Experiments on six QCIF videos with resolution \(176\times 144\) and different configuration files have been performed. Results demonstrate that the proposed technique gives enhanced video compression quality. Test videos and code used in the research is available as supplement at http://bit.ly/HEVC-NSGA-II.

Combined artificial bee colony algorithm and machine learning techniques for prediction of online consumer repurchase intention

Transactions through the web are now a progressive mechanism to access an ever-increasing range of services over more and more diverse environments. The internet provides many opportunities for companies to provide personalized online services to their customers, but the quality and novelty of some web services may adversely affect the appeal and user gratification. In the future, prediction of the consumer intention needs to be a main focus for selecting the web services for an application. The aim of this study is to predict online consumer repurchase intentions; to accomplish this objective a hybrid approach is chosen with a combination of machine learning techniques and artificial bee colony (ABC) algorithm being used. The study starts with identification of consumer characteristics for repurchase intention, followed by determining the feature selection of consumer characteristics and shopping mall attributes (with >0.1 threshold value) for the prediction model using ABC. Finally, validation using k-fold cross has been employed to measure the best classification model robustness. The classification models, viz. decision trees (C5.0), AdaBoost, random forest, support vector machine and neural network, are utilized to predict consumer purchase intention. Performance evaluation of identified models on training–testing partitions (70–30%) of the data set shows that the AdaBoost method outperforms other classification models, with sensitivity and accuracy of 0.95 and 97.58%, respectively, on testing the data set. Examining the consumer repurchase intentions by considering both shopping mall and consumer characteristics makes this study unique.

Multilevel ensemble model for prediction of IgA and IgG antibodies

Identification of antigen for inducing specific class of antibody is prime objective in peptide based vaccine designs, immunodiagnosis, and antibody productions. Its urge to introduce a reliable system with high accuracy and efficiency for prediction. In the present study, a novel multilevel ensemble model is developed for prediction of antibodies IgG and IgA. Epitope length is important in training the model and it is efficient to use variable length of epitopes. In this ensemble approach, seven different machine learning models are combined to predict variable length of epitopes (4 to 50). The proposed model of IgG specific epitopes achieves 94.43% of accuracy and IgA specific epitopes achieves 97.56% of accuracy with repeated 10-fold cross validation. The proposed model is compared with the existing system i.e. IgPred model and outcome of proposed model is improved.

Eye state prediction using ensembled machine learning models

As electric signals are transmitted between the brain cell s for transferring of data within the brain, capturing of these signals can result in understanding the functionality of brain and other directly linked parts (like eyes, ears, spinal nerves etc) of our body. We can also capture epileptic seizures that are caused by a disruption in the working of brain, by the Electro Encephalogram Test. These electric signals are to be captured by small electrodes placed on human scalp using a standard 10/20 system on an Electro Encephalograph monitor. In this work, we will predict the state of eye (open or closed) by exploring 13 machine learning models on a 15 features dataset of an EEG test. The records of 14 electrodes are used for this prediction. Results are evaluated using 6 different machine learning parameters i.e. Sensitivity, Confusion matrix, Kappa value, Specificity, Accuracy and Receiver Operating Characteristics (ROC) curve. K-fold validation and ensembling of models will be done on best three predictive models pertaining to our dataset.