scispace - formally typeset
Search or ask a question
Author

Sanjiban Sekhar Roy

Bio: Sanjiban Sekhar Roy is an academic researcher from VIT University. The author has contributed to research in topics: Rough set & Computer science. The author has an hindex of 12, co-authored 55 publications receiving 500 citations.


Papers
More filters
Book ChapterDOI
17 Jan 2017
TL;DR: The experimental results show that the accuracy of intrusion detection using Deep Neural Network is satisfactory and the potential capability of Deep Neural network as a classifier for the different types of intrusion attacks is checked.
Abstract: Security of data is considered to be one of the most important concerns in today’s world. Data is vulnerable to various types of intrusion attacks that may reduce the utility of any network or systems. Constantly changing and the complicated nature of intrusion activities on computer networks cannot be dealt with IDSs that are currently operational. Identifying and preventing such attacks is one of the most challenging tasks. Deep Learning is one of the most effective machine learning techniques which is getting popular recently. This paper checks the potential capability of Deep Neural Network as a classifier for the different types of intrusion attacks. A comparative study has also been carried out with Support Vector Machine (SVM). The experimental results show that the accuracy of intrusion detection using Deep Neural Network is satisfactory.

102 citations

Journal ArticleDOI
TL;DR: A hybrid model that uses MARS to evaluate the importance of every parameter in the prediction and these important parameters have been fed to the ELM to build hybrid model and it can be seen that this boosts the ELm performance to match up to the accuracy of MARS with lesser computation time.
Abstract: Heating load and cooling forecasting are essential for estimating energy consumption, and consequently, helping engineers in improving the energy performance right from the design phase of buildings. The capacity of heating ventilation and air-conditioning system of the building contributes to the operation cost. Moreover, building being one of the sectors with heavy energy use, it is required to develop an accurate model for energy forecasting of building and constructing energy-efficient buildings. This paper explores different machine learning techniques for predicting the heating load and cooling load of residential buildings. Among these methods, we focus on advanced techniques like Multivariate Adaptive Regression Splines (MARS), Extreme Learning Machine (ELM) and a hybrid model of MARS and ELM along with a comparison of the results with those of more conventional methods like linear regression, neural network, Gaussian processes and Radial Basis Function Network. The MARS model is a non-parametric regression model that splits the data and fits each interval into a basis function and ELM is similar to a Single Layer Feed-forward Neural Network except that in ELM randomly assigned input weights are not updated. As an improvement, we have tried a hybrid model that uses MARS to evaluate the importance of every parameter in the prediction and these important parameters have been fed to the ELM to build hybrid model and it can be seen that this boosts the ELM performance to match up to the accuracy of MARS with lesser computation time. Finally, a comparative study examines the performances of the different techniques by measuring different performance metrics.

101 citations

Book ChapterDOI
01 Jan 2015
TL;DR: A Least Absolute Shrinkage and Selection Operator (LASSO) method based on alinear regression model is proposed as a novel method to predict financial market behavior and results indicate that the proposed model outperforms the ridge linear regression model.
Abstract: Predicting stock exchange rates is receiving increasing attention and is a vital financial problem as it contributes to the development of effective strategies for stock exchange transactions. The forecasting of stock price movement in general is considered to be a thought-provoking and essential task for financial time series’ exploration. In this paper, a Least Absolute Shrinkage and Selection Operator (LASSO) method based on a linear regression model is proposed as a novel method to predict financial market behavior. LASSO method is able to produce sparse solutions and performs very well when the numbers of features are less as compared to the number of observations. Experiments were performed with Goldman Sachs Group Inc. stock to determine the efficiency of the model. The results indicate that the proposed model outperforms the ridge linear regression model.

66 citations

Journal ArticleDOI
TL;DR: The proposed model is deep neural network (DNN), which presents a category of learning algorithms that adopt nonlinear extraction of information in several steps within a hierarchical framework, primarily applied for learning and pattern classification.
Abstract: Heating load and cooling load forecasting are crucial for estimating energy consumption and improvement of energy performance during the design phase of buildings. Since the capacity of cooling ventilation and air-conditioning system of the building contributes to the operation cost, it is ideal to develop accurate models for heating and cooling load forecasting of buildings. This paper proposes a machine-learning technique for prediction of heating load and cooling load of residential buildings. The proposed model is deep neural network (DNN), which presents a category of learning algorithms that adopt nonlinear extraction of information in several steps within a hierarchical framework, primarily applied for learning and pattern classification. The output of DNN has been compared with other proposed methods such as gradient boosted machine (GBM), Gaussian process regression (GPR) and minimax probability machine regression (MPMR). To develop DNN model, the energy data set has been divided into training (70%) and testing (30%) sets. The performance of proposed model was benchmarked by statistical performance metrics such as variance accounted for (VAF), relative average absolute error (RAAE), root means absolute error (RMAE), coefficient of determination (R2), standard deviation ratio (RSR), mean absolute percentage error (MAPE), Nash–Sutcliffe coefficient (NS), root means squared error (RMSE), weighted mean absolute percent error (WMAPE) and mean absolute percentage Error (MAPE). DNN and GPR have produced best-predicted VAF for cooling load and heating load of 99.76% and 99.84% respectively.

55 citations

Journal ArticleDOI
TL;DR: The proposed architecture uses a dilated convolutional filter to obtain a larger receptive field which leads to a greater accuracy in segmenting the retinal blood vessels with near human accuracy.
Abstract: Medical diagnosis is being assisted by numerous expert systems that have been developed to increase the accuracy of such diagnoses. The development of image processing techniques along with the rapid development in areas like machine learning and computer vision help in creating such expert systems that almost nearly match the accuracy of the expert human eye. The medical condition of diabetic retinopathy is diagnosed by analyzing the retinal blood vessels for damages, abnormal new growths and ruptures. Various techniques using convolutional neural networks have been used to segment retinal blood vessels from fundus images, but these techniques often do not segment the retinal blood vessels accurately and add additional noise due to the limited receptive field of the convolutional filters. The limited receptive field of the convolutional layer prevents the convolutional neural network from getting an accurate context of objects that extend beyond the size of the filter. The proposed architecture uses a dilated convolutional filter to obtain a larger receptive field which leads to a greater accuracy in segmenting the retinal blood vessels with near human accuracy. The convolutional neural networks were trained using the popular datasets. The proposed architecture produced an area under ROC curve (AUC) of 0.9794 and an accuracy of 95.61% and required very few iterations to train the network.

39 citations


Cited by
More filters
Journal ArticleDOI
01 Feb 2020
TL;DR: A survey of deep learning approaches for cyber security intrusion detection, the datasets used, and a comparative study to evaluate the efficiency of several methods are presented.
Abstract: In this paper, we present a survey of deep learning approaches for cybersecurity intrusion detection, the datasets used, and a comparative study. Specifically, we provide a review of intrusion detection systems based on deep learning approaches. The dataset plays an important role in intrusion detection, therefore we describe 35 well-known cyber datasets and provide a classification of these datasets into seven categories; namely, network traffic-based dataset, electrical network-based dataset, internet traffic-based dataset, virtual private network-based dataset, android apps-based dataset, IoT traffic-based dataset, and internet-connected devices-based dataset. We analyze seven deep learning models including recurrent neural networks, deep neural networks, restricted Boltzmann machines, deep belief networks, convolutional neural networks, deep Boltzmann machines, and deep autoencoders. For each model, we study the performance in two categories of classification (binary and multiclass) under two new real traffic datasets, namely, the CSE-CIC-IDS2018 dataset and the Bot-IoT dataset. In addition, we use the most important performance indicators, namely, accuracy, false alarm rate, and detection rate for evaluating the efficiency of several methods.

464 citations

Journal ArticleDOI
TL;DR: A review of studies developing data-driven models for building scale applications with a focus on the input data characteristics and data pre-processing methods, the building typologies considered, the targeted energy end-uses and forecasting horizons, and accuracy assessment.

422 citations

Journal ArticleDOI
TL;DR: This survey paper describes a literature review of deep learning methods for cyber security applications, including deep autoencoders, restricted Boltzmann machines, recurrent neural networks, generative adversarial networks, and several others.
Abstract: This survey paper describes a literature review of deep learning (DL) methods for cyber security applications. A short tutorial-style description of each DL method is provided, including deep autoencoders, restricted Boltzmann machines, recurrent neural networks, generative adversarial networks, and several others. Then we discuss how each of the DL methods is used for security applications. We cover a broad array of attack types including malware, spam, insider threats, network intrusions, false data injection, and malicious domain names used by botnets.

343 citations

Journal ArticleDOI
01 Jun 2019-Heliyon
TL;DR: A systematic review of some of the popular machine learning based email spam filtering approaches and recommended deep leaning and deep adversarial learning as the future techniques that can effectively handle the menace of spam emails.

267 citations