University of Pittsburgh

Utilizing Vector Space Models for Identifying Legal Factors from Text

graduate student
Friday, November 17, 2017 - 1:00pm - 1:30pm

While a variety of knowledge representations and formalisms have been applied to legal reasoning and decision making, text mining from legal documents has been limited to general aspects such as identifying argumentation structure and textual entailment. Moreover, legal information retrieval systems often rely on hand-crafted indices to provide semantic search capabilities that are both labor-intensive and costly to develop. Vector Space Models (VSMs) of semantics emerged as an alternative approach to overcome these limitations by representing documents as points in a vector space derived from term frequencies in the corpus. This level of abstraction provides a flexible way to represent complex semantic concepts through vectors, matrices, and higher-order tensors. In this paper we utilize a number of VSMs on a corpus of judicial decisions. We focus on finding a VSM that is representative of legal factors--stereotypical fact patterns that tends to strengthen or weaken a side's argument in a legal claim. We apply different VSMs to a corpus of trade secret misappropriation cases and evaluate these representations using a document classification framework. Experimental results shows that, VSMs hold promise if applied judiciously.

Copyright 2009 | Web site by UMC Web Team