Paper: Towards Automatic Generation of Gene Summary

ACL ID W09-1312
Title Towards Automatic Generation of Gene Summary
Venue Workshop on Biomedical Natural Language Processing
Year 2009

In this paper we present an extractive system that au- tomatically generates gene summaries from the biomed- ical literature. The proposed text summarization system selects and ranks sentences from multiple MEDLINE abstracts by exploiting gene-specific information and similarity relationships between sentences. We evaluate our system on a large dataset of 7,294 human genes and 187,628 MEDLINE abstracts using Recall-Oriented Understudy for Gisting Evaluation (ROUGE), a widely used automatic evaluation metric in the text summariza- tion community. Two baseline methods are used for comparison. Experimental results show that our system significantly outperforms the other two methods with regard to all ROUGE metrics. A demo website of our system is freely accessible at http://60.19...