Description of the LIPN Systems at TAC 2008: Summarizing Information and Opinions

The Text Analysis Conferences (TAC) offer a unique occasion to show innovative approaches to text summarization. As a first incursion into this new research area, LIPN participated in the following tasks of TAC 2008: the Update Summarization task and the Opinion Summarization task. LIPN systems obtained good results for Opinion Summarization; however, we are confident that there is room for improvement, especially for the systems that did not perform so well on the Update task. This paper gives a technical description of the systems developed. Algorithms and results are then briefly discussed for the different tasks we participated in. In the first section, we describe the LIPN1-Opinion system for Opinion Summarization, which is based on an adaptation of the MEAD system (Radev et al. 2004) integrating a sentiment analysis component. We then describe a series of other systems based on an original clustering-based summarization algorithm called CBASES, which has been applied to the Update as well as to the Opinion Summarization tasks.