A Semantic Scattering Model for the Automatic Interpretation of Genitives

Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing (HLT/EMNLP), pages 891–898, Vancouver, October 2005. c⃝2005 Association for Computational Linguistics A Semantic Scattering Model for the Automatic Interpretation of Genitives Dan Moldovan Language Computer Corporation Richardson, TX 75080 moldovan@languagecomputer.com Adriana Badulescu Language Computer Corporation Richardson, TX 75080 adriana@languagecomputer.com Abstract This paper addresses the automatic clas- sification of the semantic relations ex- pressed by the English genitives. A learn- ing model is introduced based on the sta- tistical analysis of the distribution of gen- itives’ semantic relations on a large cor- pus. The semantic and contextual fea- tures of the genitive’s noun phrase con- stituents play a key role in the identifica- tion of the semantic relation. The algo- rithm was tested on a corpus of approx- imately 2,000 sentences and achieved an accuracy of 79% , far better than 44% ac- curacy obtained with C5.0, or 43% ob- tained with a Naive Bayes algorithm, or 27% accuracy with a Support Vector Ma- chines learner on the same corpus. 1 Introduction 1.1 Problem Description The identification of semantic relations in open text is at the core of Natural Language Processing and many of its applications. Detecting semantic rela- tions is useful for syntactic and semantic analysis of text and thus plays an important role in automatic text understanding and generation. Furthermore, se- mantic relations represent the core elements in the organization of lexical semantic knowledge bases used for inferences. Recently, there has been a re- newed interest in text semantics fueled in part by the complexity of some major research initiatives in Question Answering, Text Summarization, Text Understanding and others, launched in the United States and abroad. Two of the most frequently used linguistic con- structions that encode a large set of semantic rela- tions are the s-genitives, e.g. “man’s brother”, and the of-genitives, e.g. “dress of silk”. The interpreta- tion of these phrase-level constructions is paramount for various applications that make use of lexical se- mantics. Example: “The child’s mother had moved the child from a car safety seat to an area near the open passenger-side door of the car.” (The Desert Sun, Monday, October 18th, 2004). There are two semantic relations expressed by genitives: (1) “child’s mother” is an s-genitive en- coding a KINSHIP relation, and (2) “passenger-side door of the car” is an of-genitive encoding a PART- WHOLE relation. This paper provides a detailed corpus analysis of genitive constructions and a model for their auto- matic interpretation in English texts. 1.2 Semantics of Genitives In English there are two kinds of genitives. In gen- eral, in one, the modifier is morphologically linked to the possessive clitic ’s and precedes the head noun (s-genitive, i.e. NPmodif’s NPhead), and in the second one the modifier is syntactically marked by the preposition of and follows the head noun (of- genitive, i.e. NPhead of NPmodif). Although the genitive constructions have been studied for a long time in cognitive linguistics, their semantic investigation proved to be very …

Content truncated to 500 words