Dr. M S Bhargavi

"Never Stop Learning; For When We Stop Learning,

We Stop Growing"

Loyal 'Jack' Lewman

Dr. M S Bhargavi

Associate Professor

Department of Computer Science and Engineering

Bangalore Institute of Technology

Bengaluru, Karnataka, India

E-mail: ms.bhargavi@gmail.com , ms.bhargavi@ieee.org

Research Interests: Artificial Intelligence, Machine Learning, Data Mining

Educational Details:

Qualification - B.E, M.Tech, Ph.D

Bachelor of Engineering in Information Science and Engineering from Visvesvaraya Technological University (VTU), Karnataka, India in the year 2004.
Master of Technology in Software Engineering from VTU in the year 2006.
Ph.D in Computer Science under V.T.U, awarded in the year 2019

Work Experience:

Working as Associate Professor, Department of Computer Science and Engineering, Bangalore Institute of Technology, Karnataka, India since July 2022
Worked as Assistant Professor, Department of Computer Science and Engineering, Bangalore Institute of Technology, Karnataka, India between April 2009 - June 2022.
Worked as Senior Consultant Engineer, Aeronautical Development Agency (ADA), Bengaluru, Karnataka, India between March 2007 - March 2009.
Worked as Head of the Department, Information Technology, AWH College of Engineering, Calicut, Kerala between December 2006 and Feb 2007
Worked as Lecturer, Information Technology, AWH College of Engineering, Calicut, Kerala between July 2006 and Nov 2006

Courses handled for Undergraduates and Postgraduates

Machine Learning, Artificial Intelligence and Agent Technology, Big Data Analytics, Semantic Web and Social Networks, Java and J2EE, Programming in C and Data Structures, Management and Entrepreneurship, Object Oriented Concepts

Research – Ph.D (Pattern Recognition and Data Mining)

Title: Anomaly Detection in Moving Objects through Spatio-Temporal Pattern Mining

Proliferating technological sources has led to the explosive growth of data of moving objects termed spatio-temporal data. In a surveillance setup, spatio-temporal data is a source of auto detection of anomalies. Spatio-temporal data possessing characteristics of Big Data is voluminous, continuously generated, multi-dimensional and temporally ordered with a need for novel analytics to alarm a surveillance system in time on detection of anomalies. Detection of anomalies in such huge datasets must be independent of context to readily incorporate the system in any platform/environment without the need for supervised context dependent learning.

Significant Contributions:

In an unsupervised environment, two novel Cluster Validation Indices with dynamic termination for determination of true clusters is proposed.
A clustering technique in a supervised environment has been improvised with a novel, ensemble meta-algorithm to determine true clusters.
A novel approach for generation of patterns from clusters based on Regression Line is proposed.
A novel approach to interpret Slope and Intercept of Regression line and discover knowledge on the cohesiveness of clusters based on spread and density.
A novel framework to analyze the characteristic evolution of patterns over temporal phases based on the discovered knowledge on cohesiveness for detection of multi-dimensional disturbances among the instances of spatio-temporal data, independent of context.
A robust architectural framework for mining Infrequent/Rare patterns based on the characteristic evolution of patterns over temporal phases to alarm anomalies independent of context in spatio-temporal data.