Document Language Models, Query Models, and Risk Minimization for Information Retrieval

TLDR

The framework extends recent language‑modeling IR developments and introduces a Markov‑chain method for query modeling that connects to link‑analysis and social‑network algorithms. The authors aim to combine document and query language models into a Bayesian decision‑theoretic ranking function, proposing the Markov‑chain approach to estimate query models. They estimate document and query language models, cast retrieval as risk minimization, and evaluate the Markov‑chain method on TREC collections against basic language‑model and vector‑space baselines with Rocchio expansion. The method yields significant gains over standard query‑expansion techniques for strong TF‑IDF baselines, especially for short Web queries.

Abstract

We present a framework for information retrieval that combines document models and query models using a probabilistic ranking function based on Bayesian decision theory. The framework suggests an operational retrieval model that extends recent developments in the language modeling approach to information retrieval. A language model for each document is estimated, as well as a language model for each query, and the retrieval problem is cast in terms of risk minimization. The query language model can be exploited to model user preferences, the context of a query, synonomy and word senses. While recent work has incorporated word translation models for this purpose, we introduce a new method using Markov chains defined on a set of documents to estimate the query models. The Markov chain method has connections to algorithms from link analysis and social networks. The new approach is evaluated on TREC collections and compared to the basic language modeling approach and vector space models together with query expansion using Rocchio. Significant improvements are obtained over standard query expansion methods for strong baseline TF-IDF systems, with the greatest improvements attained for short queries on Web data.