论文信息 - Headline extraction based on a combination of uni- and multidocument summarization techniques

Headline extraction based on a combination of uni- and multidocument summarization techniques

The TNO system for multi-document summarisation is based on an extraction approach. For headline generation, we chose to extend our system to extract the most informative topical noun phrase. The cluster topic is defined as the most frequent term occurring in the most salient document sentences. The core of our system is a probabilistic model, which estimates the log-odds of salience based on a number of features including sentence position, sentence length, cue phrases and a language model based content score. The parameters of the model were estimated on annotated training data.

Wessel Kraaij | Anette Hulth | Martijn Spitters

[1] Dragomir R. Radev,et al. Centroid-based summarization of multiple documents: sentence extraction, utility-based evaluation, and user studies , 2000, ArXiv.

[2] Wessel Kraaij,et al. Using language models for tracking events of interest over time , 2001 .

[3] H. P. Edmundson,et al. New Methods in Automatic Extracting , 1969, JACM.

[4] Wessel Kraaij,et al. Twenty-One: Cross-Language Disclosure and Retrieval of Multimedia Documents on Sustainable Development , 1998, Comput. Networks.

[5] Djoerd Hiemstra,et al. Twenty-One at CLEF-2000: Translation Resources, Merging Strategies and Relevance Feedback , 2000, CLEF.

[6] Francine Chen,et al. A trainable document summarizer , 1995, SIGIR '95.

[7] Jade Goldstein-Stewart,et al. The use of MMR, diversity-based reranking for reordering documents and producing summaries , 1998, SIGIR '98.

[8] Wessel Kraaij,et al. Combining a mixture language model and Naive Bayes for multi-document summarisation , 2001 .

[9] Djoerd Hiemstra,et al. Twenty-One at TREC-8: using Language Technology for Information Retrieval , 1999, TREC.