Query expansion using local and global document analysis

By Jinxi Xu, W. Bruce Croft. 1996

Automatic query expansion has long been suggested as a technique for dealing with the fundamental issue of word mismatch in information retrieval. A number of approaches to ezpanrnion have been studied and, more recently, attention has focused on techniques that analyze the corpus to discover word relationship (global techniques) and those that analyze documents retrieved by the initial quer~ ( local feedback). In this paper, we compare the effectiveness of these approaches and show that, although global analysis haa some advantages, local analysia is generally more effective. We also show that using global analysis techniques, such as word contezt and phrase structure, on the local aet of documents produces re-sults that are both more effective and more predictable than simple local feedback.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Relevance-based Word Embedding (2017, background): “As is widely known in the information retrieval literature [11, 38], there is a big difference between the unigram distribution of words on sub-topics of a collection and the unigram distribution estimated from the whole collection.”
Task-Oriented Query Reformulation with Reinforcement Learning (2017, background): “It typically expands the original query by adding terms from, for instance, dictionaries of synonyms such as WordNet (Miller, 1995), or from the initial set of retrieved documents (Xu and Croft, 1996).”
Piggyback: Using Search Engines for Robust Cross-Domain Named Entity Recognition (2011, background): “Perhaps the oldest and best known application is pseudo-relevance feedback which uses words and phrases from search results for query expansion (Rocchio, 1971; Xu and Croft, 1996).”
Estimating Embedding Vectors for Queries (2016, methodology): “As suggested by Xu and Croft [36], query expansion techniques can use global or local document analysis.”
Retrieval-Enhanced Machine Learning (2022, background): “It has shown successful results in a wide range of retrieval tasks [2, 8, 14, 28, 36, 68, 73], demonstrat-”
From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing (2018, background): “This is also known as a local approach for query expansion [50].”
Regularized estimation of mixture models for robust pseudo-relevance feedback (2006, background): “The basic idea of pseudo feedback is to assume a certain number of top-ranked documents are relevant and learn from these documents to improve the query representation, thus to improve retrieval accuracy [1, 16].”
Embedding-based Query Language Models (2016, methodology): “[1] N. Abdul-jaleel, J. Allan, W. B. Croft, F. Diaz, L. Larkey, X. Li, D. Metzler, M. D. Smucker, T. Strohman, H. Turtle, and C. Wade.”
Personalized query expansion for the web (2007, background): “RF has also been shown to be effectively automatized by considering the top ranked documents as relevant [37] (this is known as Pseudo RF).”
Relevant term suggestion in interactive web search based on contextual information in query session logs (2003, background): “Information retrieval systems provide the desired function by invoking an interactive process of query reformulation and relevance feedback (Efthimiadis, 1996; Spink & Losee, 1996; Xu & Croft, 1996).”
Bridging the lexical chasm: statistical approaches to answer-finding (2000, background): “Synonymy has received much attention within the document retrieval community recently, and researchers have applied a variety of heuristic and statistical techniques|including pseudorelevance feedback, local context analysis, and probabilistic models of synonymy [7; 13; 2].”
Query Expansion Techniques for Information Retrieval: a Survey (2017, unknown): “This approach is known as local context analysis [277].”
Learning to Reweight Terms with Distributed Representations (2015, background): “Recently, Zhao et al. [22] proposed a framework to construct features for query terms from pseudo relevance feedback [20] and use them to predict term weight.”