Supporting analysis of future-related information in news archives and the web

A lot of future-related information is available in news articles or Web pages. This information can however differ to large extent and may fluctuate over time. It is therefore difficult for users to manually compare and aggregate it, and to re-construct the most probable course of future events. In this paper we approach a problem of automatically generating summaries of future events related to queries using data obtained from news archive collections or from the Web. We propose two methods, explicit and implicit future-related information detection. The former is based on analyzing the context of future temporal expressions in documents, while the latter relies on detecting periodical patterns in historical document collections. We present a graph-based visualization of future-related information and demonstrate its usefulness through several examples. Categories and Subject Descriptors H.3.3 [Information Storage and Retrieval]: Information Search and Retrieval General Terms Algori...

• Type: Conference
• Year: 2009
• Where: JCDL
• Authors: Adam Jatowt, Kensuke Kanazawa, Satoshi Oyama, Katsumi Tanaka
• Keywords: Education, Future Events, Future Temporal Expressions, Future-related Information, JCDL 2009