Downloads: 110 | Views: 196
M.Tech / M.E / PhD Thesis | Computer Science & Engineering | India | Volume 4 Issue 6, June 2015
Identifying Emerging Topics Using Link Anomaly Detection in Social Media
Shweta V. Saswade | Prof. S. S. Nandgaonkar [2]
Abstract: In data mining era, detecting and generating new concepts has attracted much attention. To discover, the emergence of new topics in news data is a biggest challenge in data mining. This problem can be enlarged as discovering a new concept. Few years ago, domain experts detected emergence of new stories. But it is very critical and time consuming task to read stories and concluding misbehaviors manually. In addition, mapping these misbehaviors to various stories requires excellent knowledge about the old concepts and news. Also automatically modeling a new concept has much importance in data mining. The outliers in news are the basic clues for concluding the emergence of a new story. The outliers are the s which doesn-t match the whole concept of the news. These outliers are mapped to the stories where these s does not behave as outliers. After mapping these outliers, anomaly linking can generate a new concept which can be modeled as emerging story. News Classification, Anomaly Detection, Concept Detection and Generation techniques can be used to efficiently model the new concept.
Keywords: News Classification, Anomaly Detection, Anomaly Linking, Concept Detection, Concept Generation
Edition: Volume 4 Issue 6, June 2015,
Pages: 1676 - 1682