Query Expansion in Text Information Retrieval with Local Context and Distributional Model

Abstract

The Semantic Distributional Model is based on the frequency of contexts of use of language terms in large open corpus such as the web, to establish similarity or the relationship between words. These relationships or similarities can be used to add terms when expanding queries. The idea explored in this paper is that, for queries in closed collections of text documents, a posterior filter based on the restricted vocabulary of the collection can improve the effectiveness of automatic query expansion. This idea is developed and evaluated in publicly available benchmarks presenting promising results. Subject Categories and Descriptors: [H.3.3 Information Search and Retrieval]; Query formulation: [I.2.7 Natural Language Processing] Text analysis [F.4.2 Grammars and Other Rewriting Systems]; Grammar types General Terms: Distributional Semantic Model, Information Retrieval, Local Context Analysis.