A Survey on Anomaly Detection for Discovering Emerging Topics?
Journal: International Journal of Computer Science and Mobile Computing - IJCSMC (Vol.3, No. 10)Publication Date: 2014-10-30
Authors : S.Saranya; R.Rajeshkumar; S.Shanthi;
Page : 895-902
Keywords : Topic detection; temporal text mining; burst detection; anomaly detection;
Abstract
This paper identifies various concepts involved in social networks for finding the emerging topics. We focus on the various methods that can be applied for detecting the anomaly. The methods used are Hidden Markov Model, UMass Approach, CMU Approach, Change Finder method and Finite Mixture Model. These methods involve texts, videos, audios, URLs and mentions which are shared in the social networks. Kullback- Leibler divergence measure is used here to discover coherent themes and topics over time.
Other Latest Articles
Last modified: 2014-11-03 21:47:35