KEYWORD AND NAME ENTITY IDENTIFICATION BASED NEWS TOPIC THREAD EXTRACTION
Sjtu Joint · Computer Applications and Software · 2011
How to automatically organize news topic,understand news topic from different angles and aspects to solve the problem of network news information overloading is becoming a hotspot for research.The article introduces a news topic threading view,i.e.,according to the thread extraction algorithm,Keywords and name entity aggregations are obtained as each thread theme,then classify news reports into threads as contents to organize the news topic.Experiments indicate that the method proposed in the article performs good at thread accuracy and document dividing assessment indices and can moderately clearly demonstrate different threads of topic to help users understand the news topic development skeleton.