Summaries on the fly: Query-based extraction of structured knowledge from web documents

Abstract

A large part of Web resources consists of unstructured textual content. Processing and retrieving relevant content for a particular information need is challenging for both machines and humans. While information retrieval techniques provide methods for detecting suitable resources for a particular query, information extraction techniques enable the extraction of structured data and text summarization allows the detection of important sentences. However, these techniques usually do not consider particular user interests and information needs. In this paper, we present a novel method to automatically generate structured summaries from user queries that uses POS patterns to identify relevant statements and entities in a certain context. Finally, we evaluate our work using the publicly available New York Times corpus, which shows the applicability of our method and the advantages over previous works. The final publication is available at Springer via https://doi.org/10.1007/978-3-642-39200-9_22