Thomson Reuters at TAC 2008: Aggressive Filtering with FastSum for Update and Opinion Summarization
نویسندگان
چکیده
In TAC 2008 we participated in the main task (Update Summarization) as well as the Sentiment Summarization pilot task. We modified the FastSum system (Schilder and Kondadadi, 2008) and added more aggressive filtering in order to adapt the system to update summarization and sentiment summarization. For the Update Summarization task, we show that a classifier that identifies sentences that are similar to typical first sentences of a news article improves the overall linguistic quality of the generated summaries. For the Sentiment Summarization pilot task, we use a simple sentiment classifier based on a gazetteer of positive and negative sentiment words derived from the General Inquirer and other sources to produce opinion-based summaries for a collection of blog posts given a set of positive and negative questions.
منابع مشابه
Thomson Reuters at TAC 2009: ContextChain and Fractional Conditional Compressibility of Models
This paper contains the results for the FastSum system and a simple baseline system for the TAC 2009 main task – update summarization –. For the pilot task of Automatically Evaluating Summaries of Peers (AESOP), we present two novel metrics. The first metric called ContextChain is an extension of a recently proposed metric AutoSummENG that is based on comparing n-gram graphs of the model summar...
متن کاملDescription 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 imp...
متن کاملUsing Coreference Links and Sentence Compression in Graph-based Summarization
Recent years have shown that graphs are an adequate text representation model for summarization. For this years’ TAC update summarization challenge, we extended our graph-based summarization system with coreference relations and sentence compression. Our results show that using coreference relations did not result in a significant performance gain; sentence compression had a negative effect on ...
متن کاملICTCAS's ICTGrasper at TAC 2008: Summarizing Dynamic Information with Signature Terms Based Content Filtering
This paper presents our new, topic-oriented multi-document summarization system used in TAC 2008. To deal with the problem of summarizing changes of the dynamic information with time going, we propose a novel summarization method with signature terms based content filtering. We first present the definition of dynamic summarization according to temporal analysis and then propose the fundamental ...
متن کاملNUS at TAC 2008: Augumenting Timestamped Graphs with Event Information and Selectively Expanding Opinion Contexts
Our team explored several new approaches in the update and opinion summarization tasks in the Text Analysis Conference (TAC) 2008. For the update task, we refine our previous timestamped graph approach by incorporating information about temporal ordering of events in the articles, by using the publicly available Tarsqi toolkit. For the pilot opinion task, we utilize the provided opinion snippet...
متن کامل