Discourse Parsing: A Decision Tree Approach

The paper presents a new statistical method, for parsing discourse. A parse of discourse is defined as a set of semantic dependencies among sentences tha t make up the discourse. A collection of news articles from a Japanese economics daily are manually marked for dependency and used as a training/ test ing corpus. We use a C4.5 decision tree method to develop a model of sentential dependencies. However, rather than to use class decisions made by C4.5, we exploit information on class distr ibutions to rank possible dependencies among sentences according to their probabilistic s t rength and take a parse to be a set of highest ranking dependencies. We also study effects of features such as clue words, distance and similarity on the performance of the discourse parser. Experiments have found tha t the method performs reasonably well on diverse text types, scoring an accuracy rate of over 60%. 1 I n t r o d u c t i o n At tempts to the automatic identification of a structure in discourse have so far met with a limited success in the computat ional linguistics li terature. Par t of the reason is that, compared to sizable da ta resources available to parsing research such as the Penn Treebank (Marcus et al., 1993), large corpora annotated for discourse information are hard to come by. Researchers in discourse usually work with a corpus of a few hundred sentences (Kurohashi and Nagao, 1994; Li tman and Passonneau, 1995; Hearst, 1994). The lack of a large-scale corpus has made it impossible to talk about results of discourse studies with the sufficient degree of reliability. In the work described here, we created a corpus with discourse information, containing 645 articles from a Japanese economic paper, an order of magnitude larger than any previous work on discourse processing. It had a total of 12.770 sentences and 5,352 paragraphs. Each article in the corpus was manually annotated for a discourse dependency" relation. We then built a statistical discourse parser based on the C4.5 decision tree method (Quinlan, 1993), which was trained and tested on the corpus we have creFigure 1: A discourse tree. 'S' denotes a sentence. ated. The design of a parser was inspired by Haruno (1997)'s work on statistical sentence parsing. The paper is organized as follows. Section 2 presents general ideas about statistical parsing as applied to the discourse, After a brief introduction to some of the points of a decision tree model, we discuss incorporating a decision tree within a statistical parsing model. In Section 3, we explain how we have built an annotated corpus. There we also describe a procedure of experiments we have conducted, and conclude the section with their results. 2 S t a t i s t i c a l D i s c o u r s e P a r s i n g First, let us make ourselves clear about what we mean by parsing a discourse. The job of parsing is to find whatever dependencies there are among elements that make up a part icular linguistic unit. In discourse parsing, elements one is interested in finding dependencies among correspond to sentences, and a level of unit under investigation is a discourse. We take a naive approach to the notion of a dependency here. We think of it as a re!ationship between a pair of sentences such that the interpretat ion of one sentence in some way depends on that of the other. Thus a dependency relationship is not a s t ructural one, but rather a semantic or rhetorical one. The job of a discourse parser is to take as input