Department of Computer Science and Information Systems

Permanent URI for this collectionhttp://localhost:4000/handle/123456789/1928

Browse

End date: 2019Subject: search.filters.subject.Machine Learning

Search Results

Now showing 1 - 10 of 16
  • No Thumbnail Available
    Item
    Performance Analysis of Machine Learning Algorithms for Fall Detection
    (IEEE, 2019) Ramachandran, Anita
    Intelligent IoT-based ambient assisted living systems (AALS) have been a major research focus area in recent times. Application of machine learning in areas of AALS such as fall detection has the potential to have huge social impact. There has been active research in the application of machine learning in fall detection, using data generated by various means such as wearable devices, environment sensors and vision based systems. The main challenge is to create a model that detects falls accurately, while keeping the design of the fall detection system minimal and non-intrusive. Wearable devices equipped with inertial motion unit (IMU) sensors and vital signs sensors are commonly used to enable analysis around performance of machine learning (ML) models. In this paper, we analyze the impact of using IMU sensor parameters in combination with vital signs parameters, on the performance of ML algorithms for fall detection. We present details on the data set we have generated for this purpose, and compare the performance of various ML algorithms on the collected dataset, with features from IMU sensors vis-à-vis those from IMU sensors in combination with vital signs sensors. We also apply machine learning algorithms on two public datasets, one with only IMU sensor parameter values and the second with only vital signs parameter values, and summarize their performance.
  • No Thumbnail Available
    Item
    Machine Learning-based Fall Detection in Geriatric Healthcare Systems
    (IEEE, 2018) Ramachandran, Anita
    Intelligent IoT-based ambient assisted living systems (AALS) have been a major research focus area in recent times. According to the studies conducted by the Govt. of India, elderly population in India has reached 8.3% of the total population [40]. Per the National Program for Health Care of the Elderly (NPHCE), the elderly population in India has tripled over the last 50 years, and is projected to increase to 33.32 million by 2021 and 300.96 million by 2051 [41]. Application of machine learning in AALS, such as fall detection, therefore, has the potential to have huge public impact. In this paper, we propose a fall detection system that takes into account not only various wearable sensor node parameter readings for a subject, but also his biological and physiological profile. The profile is used to determine a fall risk category for the subject. We performed machine learning experiments using public datasets for fall detection which included wearable sensor node readings. The algorithms were then retrained by feeding in the risk categorization of the subject, and results from this analyses are presented. The objective of the experiments was to find out the impact of a subject's risk categorization on the accuracy of fall detection. The algorithms presented here form part of a comprehensive geriatric healthcare system under development, which comprises wearable sensor nodes, coordinator nodes, an indoor localization framework and cloud-hosted application servers. A brief overview of the system capabilities is also presented.
  • No Thumbnail Available
    Item
    Machine Learning-Based Techniques for Fall Detection in Geriatric Healthcare Systems
    (IEEE, 2018) Ramachandran, Anita
    Intelligent IoT-based ambient assisted living systems (AALS) have been a major research focus area in recent times. According to the studies conducted by National Program for Health Care of the Elderly (NPHCE), elderly population in India will increase to 12% of the national population by 2025 with 8%-10% requiring utmost care. Application of machine learning in areas of AALS such as fall detection, therefore, has the potential to have huge public impact. In this paper, we propose a fall detection system that takes into account not only various wearable sensor node parameter readings for a subject, but also his biological and physiological profile. The profile is used to determine a fall risk category for the subject. We performed machine learning experiments using public datasets for fall detection which included wearable sensor node readings. The algorithms were then retrained by feeding in the risk categorization of the subject, and results from this analyses are presented. The objective of the experiments was to find out the impact of a subject's risk categorization on the accuracy of fall detection. The algorithms presented here form part of a comprehensive geriatric healthcare system under development, which comprises wearable sensor nodes, co-ordinator nodes, an indoor localization framework and cloud-hosted application servers. A brief overview of the system capabilities is also presented.
  • No Thumbnail Available
    Item
    PACE: Platform for Android Malware Classification and Performance Evaluation
    (IEEE, 2019) Agarwal, Vinti
    Android malware has become the topmost threat for ubiquitous and useful Android eco-system. Multiple solutions leveraging big data and machine learning capabilities to detect android malware are being constantly developed. Too often, many of these solutions are either limited to the research output or remain isolated and unable to reach to end-users or malware researchers. In this paper, we propose, PACE, a unified solution to offer open and easy implementation access to several machine learning-based Android malware detection techniques that make most of the research in this domain reproducible. The benefits of PACE are offered using three interfaces i.e. through REST API, Web Interface and ADB interface. Multiple interfaces enable users with different expertise such as IT administrator, security practitioners, malware researcher, etc. to avail its offered services. A community-accepted dataset is used for testing of all the techniques to provide a better comparison of performance. A prototype of the proposed platform is introduced and our vision is that it will help malware analysts to tackle challenges and reduce the amount of manual work.
  • No Thumbnail Available
    Item
    Predicting the dynamics of social circles in ego networks using pattern analysis and GA K-means clustering
    (Wiley, 2015-04) Agarwal, Vinti
    The tremendous amount of content generated on online social networks has led to a radical paradigm shift in the interest of people to group friends dynamically and share content selectively. At large, social networking sites (e.g. Google+, Facebook, Twitter, etc.) offer users with various controls over categorizing their family members, friends, colleagues, etc. into one or more ‘circles’ that they want to share content with. However, it is typically impossible to design social circles in large networks and update their number and size, whenever networks grow. Aiming at predicting the dynamics of formation and evolution of social circles, we performed several experiments on ground-truth data, and found that studying patterns of network and profile features at individual level rather than studying circle as a whole can greatly enhance the understanding of social circles development in online social networks. In this review, we first present a comprehensive study of the structural behavior of circles, and then build an observation that within every circle there exist some key elements, termed as ‘Node of Creations (NoCs)’, which play an important role in the development, survival, and evolvability of circle structures. We, therefore, propose a Genetic Algorithm–based framework to determine these key elements (NoCs) and differentiate Ego networks into non-overlapping, hierarchically nested as well as overlapping circles by leveraging knowledge from the identified patterns in order to assist K-means clustering. We have performed our experiments using Facebook and Twitter datasets and the experimental results clearly demonstrate the effectiveness of our scheme. WIREs Data Mining Knowl Discov 2015, 5:113–141. doi: 10.1002/widm.1150
  • No Thumbnail Available
    Item
    Feature Selection for Detection of Peer-to-Peer Botnet Traffic
    (ACM Digital Library, 2013) Narang, Pratik
    The use of anomaly-based classification of intrusions has increased significantly for Intrusion Detection Systems. Large number of training data samples and a good ‘feature set’ are two primary requirements to build effective classification models with machine learning algorithms. Since the amount of data available for malicious traffic will often be small compared to the available traces of benign traffic, extraction of ‘good’ features which enable detection of malicious traffic is a challenging area of work. This research work presents preliminary results of comparison of performance of three different feature selection algorithms - Correlation based feature selection, Consistency based subset evaluation and Principal component analysison three different Machine learning techniques- namely Decision trees, Na¨ıve Bayes classifier, and Bayesian Network classifier. These algorithms are evaluated for the detection of Peer-to-Peer (P2P) based botnet traffic.
  • No Thumbnail Available
    Item
    PeerShark: Detecting Peer-to-Peer Botnets by Tracking Conversations
    (IEEE, 2014) Narang, Pratik
    The decentralized nature of Peer-to-Peer (P2P) botnets makes them difficult to detect. Their distributed nature also exhibits resilience against take-down attempts. Moreover, smarter bots are stealthy in their communication patterns, and elude the standard discovery techniques which look for anomalous network or communication behavior. In this paper, we propose Peer Shark, a novel methodology to detect P2P botnet traffic and differentiate it from benign P2P traffic in a network. Instead of the traditional 5-tuple 'flow-based' detection approach, we use a 2-tuple 'conversation-based' approach which is port-oblivious, protocol-oblivious and does not require Deep Packet Inspection. Peer Shark could also classify different P2P applications with an accuracy of more than 95%.
  • No Thumbnail Available
    Item
    Temperature compensation of ISFET based pH sensor using artificial neural networks
    (IEEE, 2017) Narang, Pratik; Ajmera, Pawan K.
    This paper presents a new Machine Learning based temperature compensation technique for Ion-Sensitive Field-Effect Transistor (ISFET). The circuit models for various electronic devices like MOSFET are available in commercial Technology Computer Aided Design (TCAD) tools such as LT-SPICE but no built-in model exists for ISFET. Considering SiO 2 as the sensing film, an ISFET circuit model was created in LT-SPICE and simulations were carried out to obtain characteristic curves for SiO 2 based ISFET. A Machine Learning (ML) model was trained using the data collected from the simulations performed using the ISFET macromodel in the read-out circuitry. The simulations were performed at various temperatures and the temperature drift behavior of ISFET was fed into the ML model. Constant pH (predicted by the system) curves were obtained when the device is tested for various pH (7 and 10) solutions at different ambient temperatures.
  • No Thumbnail Available
    Item
    Multiclass Fake News Detection using Ensemble Machine Learning
    (IEEE, 2019) Narang, Pratik
    Over the past few years, fake news and its influence have become a growing cause of concern in terms of debate and public discussions. Due to the availability of the Internet, a lot of user-generated content is produced across the globe in a single day using various social media platforms. Nowadays, it has become very easy to create fake news and propagate it worldwide within a short period of time. Despite receiving significant attention in the research community, fake news detection did not improve significantly due to insufficient context-specific news data. Most of the researchers have analyzed the fake news problem as a binary classification problem, but many more prediction classes exist. In this research work, experiments have been conducted using a tree-based Ensemble Machine Learning framework (Gradient Boosting) with optimized parameters combining content and context level features for fake news detection. Recently, adaptive boosting methods for classification problems have been derived as gradient descent algorithms. This formulation justifies key elements and parameters in the methods, which are chosen to optimize a single common objective function. Experiments are conducted using a multi-class dataset (FNC) and various machine learning models are used for classification. Experimental results demonstrate the effectiveness of the ensemble framework compared to existing benchmark results. Using the Gradient Boosting algorithm (an ensemble machine learning framework), we achieved an accuracy of 86% for multi-class classification of fake news having four classes.
  • No Thumbnail Available
    Item
    PeerShark: flow-clustering and conversation-generation for malicious peer-to-peer traffic identification
    (Springer, 2014-10) Narang, Pratik
    The distributed and decentralized nature of peer-to-peer (P2P) networks has offered a lucrative alternative to bot-masters to build botnets. P2P botnets are not prone to any single point of failure and have been proven to be highly resilient against takedown attempts. Moreover, smarter bots are stealthy in their communication patterns and elude the standard discovery techniques which look for anomalous network or communication behavior. In this paper, we present a methodology to detect P2P botnet traffic and differentiate it from benign P2P traffic in a network. Our approach neither assumes the availability of any ‘seed’ information of bots nor relies on deep packet inspection. It aims to detect the stealthy behavior of P2P botnets. That is, we aim to detect P2P botnets when they lie dormant (to evade detection by intrusion detection systems) or while they perform malicious activities (spamming, password stealing, etc.) in a manner which is not observable to a network administrator. Our approach PeerShark combines the benefits of flow-based and conversation-based approaches with a two-tier architecture, and addresses the limitations of these approaches. By extracting statistical features from the network traces of P2P applications and botnets, we build supervised machine learning models which can accurately differentiate between benign P2P applications and P2P botnets. PeerShark could also detect unknown P2P botnet traffic with high accuracy.