Identifying the intent of a user query using support vector machines

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

17 Scopus citations

Abstract

In this paper we introduce a high-precision query classification method to identify the intent of a user query given that it has been seen in the past based on informational, navigational, and transactional categorization. We propose using three vector representations of queries which, using support vector machines, allow past queries to be classified by user's intents. The queries have been represented as vectors using two factors drawn from click-through data: the time users take to review the documents they select and the popularity (quantity of preferences) of the selected documents. Experimental results show that time is the factor that yields higher precision in classification. The experiments shown in this work illustrate that the proposed classifiers can effectively identify the intent of past queries with high-precision.

Original languageEnglish
Title of host publicationString Processing and Information Retrieval - 16th International Symposium, SPIRE 2009, Proceedings
Pages131-142
Number of pages12
DOIs
StatePublished - 9 Nov 2009
Event16th International Symposium on String Processing and Information Retrieval, SPIRE 2009 - Saariselka, Finland
Duration: 25 Aug 200927 Aug 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5721 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Symposium on String Processing and Information Retrieval, SPIRE 2009
CountryFinland
CitySaariselka
Period25/08/0927/08/09

Fingerprint Dive into the research topics of 'Identifying the intent of a user query using support vector machines'. Together they form a unique fingerprint.

Cite this