Exploiting E-mail structure to improve summarization

Abstract

For this thesis, I designed and implemented a system to summarize e-mail messages. The system exploits two aspects of e-mail, thread reply chains and commonly-found features, to generate summaries. The system uses existing software designed to summarize single text documents. Such software typically performs best on well-authored, formal documents. E-mail messages, however, are typically neither well-authored, nor formal. As a result, existing summarization software typically gives a poor summary of e-mail messages. To remedy this poor performance, the system's approach preprocesses e-mail messages to synthesize new input to this software, so that it will output more useful summaries of e-mail. This pre-processing involves a lightweight, heuristics-based approach to filtering e-mail to remove e-mail signatures, header fields, and quoted parent messages. I also present a heuristics-based approach to identifying and reporting names, dates, and companies found in e-mail messages. Lastly, I discuss conclusions from a pilot user study of my summarization system, and conclude with areas for further investigation.