Topic Segmentation with Hybrid Document Indexing

Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, pp. 351–359, Prague, June 2007. c⃝2007 Association for Computational Linguistics Topic Segmentation with Hybrid Document Indexing Irina Matveeva Department of Computer Science University of Chicago Chicago, IL 60637 matveeva@cs.uchicago.edu Gina-Anne Levow Department of Computer Science University of Chicago Chicago, IL 60637 levow@cs.uchicago.edu Abstract We present a domain-independent unsuper- vised topic segmentation approach based on hybrid document indexing. Lexical chains have been successfully employed to evalu- ate lexical cohesion of text segments and to predict topic boundaries. Our approach is based in the notion of semantic cohesion. It uses spectral embedding to estimate seman- tic association between content nouns over a span of multiple text segments. Our method significantly outperforms the baseline on the topic segmentation task and achieves perfor- mance comparable to state-of-the-art meth- ods that incorporate domain specific infor- mation. 1 Introduction The goal of topic segmentation is to discover story boundaries in the stream of text or audio recordings. Story is broadly defined as segment of text contain- ing topically related sentences. In particular, the task may require segmenting a stream of broadcast news, addressed by the Topic Detection and Track- ing (TDT) evaluation project (Wayne, 2000; Allan, 2002). In this case topically related sentences belong to the same news story. While we are considering TDT data sets in this paper, we would like to pose the problem more broadly and consider a domain- independent approach to topic segmentation. Previous research on topic segmentation has shown that lexical coherence is a reliable indicator of topical relatedness. Therefore, many approaches have concentrated on different ways of estimating lexical coherence of text segments, such as seman- tic similarity between words (Kozima, 1993), sim- ilarity between blocks of text (Hearst, 1994), and adaptive language models (Beeferman et al., 1999). These approaches use word repetitions to evaluate coherence. Since the sentences covering the same story represent a coherent discourse segment, they typically contain the same or related words. Re- peated words build lexical chains that are conse- quently used to estimate lexical coherence. This can be done either by analyzing the number of overlap- ping lexical chains (Hearst, 1994) or by building a short-range and long-range language model (Beefer- man et al., 1999). More recently, topic segmentation with lexical chains has been successfully applied to segmentation of news stories, multi-party conversa- tion and audio recordings (Galley et al., 2003). When the task is to segment long streams of text containing stories which may continue at a later point in time, for example developing news stories, building of lexical chains becomes intricate. In ad- dition, the word repetitions do not account for syn- onymy and semantic relatedness between words and therefore may not be able to discover coherence of segments with little word overlap. Our approach aims at discovering semantic relat- edness beyond word repetition. It is based on the notion of semantic cohesion rather than lexical cohe- sion. We propose to use a similarity …

Content truncated to 500 words