Anomaly Detection in Big Data

<strong>IEEE STUDENT BRANCH JAWAHARLAL NEHRU UNIVERSITY</strong> a Special lecture on <strong>Anomaly Detection in Big Data</strong> By <strong>Mr. Chandresh Kumar Maurya</strong> Research Scholar at Indian Institute of Technology, Roorkee, India <strong>Abstract :</strong> Anomaly detection is an important data mining, machine learning task. The main focus in anomaly detection is to discover unusual patterns in the data. With the advent of Big Data many new challenges are emerging in this field, which need to be addressed by the next generation of the anomaly detection algorithms. Anomalies are of various kind such as point anomaly, contextual anomaly, subsequence anomaly etc. In this talk speaker will focus on point anomalies in Big Data. A point anomaly is a point of usually high or low value with respect to other instances. A point anomaly usually appears as a rare event and it can be modeled via class-imbalance learning approach. In this talk he will show how we can solve this class-imbalance learning problem from Cost Sensitive Learning based method in a distributed and large scale sparse learning setting. <strong>Brief Biography: </strong>Mr. Chandresh Kumar Maurya is a Ph.D. student at Indian Institute of Technology, Roorkee, India. He will be joining IBM Research at Bangalore starting from Jan 2017 as Post Doc fellow. He completed his Master's degree from JNU, New Delhi in 2012 and Bachelor's degree from BIET Jhansi in 2010. He is recipient of Prime Minister's Fellowship for Doctoral Research, awarded by Confederation of Indian Industries and Robert Bosch, Bangalore. His paper on anomaly detection accepted in ICML 2016 workshop. He won the Best Poster paper award at Xerox Research Center India (XRCI Open 2016) and Interned at Microsoft India (R&amp;D), Bangalore and Research &amp; Technology Center (RTC), Robert Bosch, Bangalore, India, in 2014. Date: <strong>November 10, 2016</strong>