A Survey of Automatic Query Expansion in Information Retrieval
ACM Computing Surveys (CSUR)
Enterprise information integration – a semantic approach
From Integrated Publication and Information Systems to Virtual Information and Knowledge Environments
Hi-index | 0.00 |
In e-commerce it is often crucial to provide customersa large choice of relevant offers. As short and incompletequeries prevail, appropriate expansion methods aredesirable. This work describes a method for the automaticconstruction of a thesaurus based on existing categoriesof documents. A clustering algorithm, the "Layer-Seedsmethod", is introduced, which facilitates the automaticgeneration of thesaurus reflecting the specific vocabularyoccurring in a given collection of documents. Theclustering works on terms extracted from the documentsin a certain category and organizes them in a tree-likehierarchical structure 驴 a thesaurus. The thesaurus isthen employed for automatic query expansion in an e-commerceapplication in order to obtain better results forproduct searching. Experiments yield evidence that asignificant increase of user satisfaction is achieved.