15-10-2014, 04:39 PM
My task was to implement the Probabilistic Mixture Model ,as described in the paper "A Probabilistic Approach For Reterospective News Event Detection " written by Zhiwei Li(Microsoft, China), Bin Wang (Microsoft ,China ) ,Mingjing Li, WeiYing Ma (Microsoft ,China).I carried out extensive data processing activities ,parsing heterogenous RSS and Atom feeds and storing the parsed feeds in relational database. I independently identified many potential approaches for improving the integrity of keywords included in the vector space model : for example ,by using N-Grams and combining OverLapping terms.