Text Summarization via Latent Semantic Analysis and Anaphora Resolution
The information-overloading problem can be reduced by automatic summarization. I present the state of the art in this area. Special attention is paid to evaluationmethods. Then I describe principles of Latent Semantic Analysis (LSA) and basics of Anaphora Resolution (AR). Further, I discuss a new summarization approach that uses LSA and AR. A summary evaluation method based on LSA is proposed as well. Finally, I show preliminary results and discuss further research directions. The final goal is to integrate the summarizer into a natural language processing system capable of searching and presenting documents in a concise and coherent form.
Keywords: Summarization, latent semantic analysis, anaphora resolution, sentence compression, summary evaluation, multilingual searching
Year: 2005
Authors of this publication:
Josef Steinberger
E-mail: jstein@kiv.zcu.cz