A Word Similarity Based Belief Network IR Model with Two Term Layers

The retrieval performance of an information system usually increases when it uses the relationships among the terms in a given document collection. To resolve the problem that Belief Network model does not use term relationships, this paper presents an extended model with two term layers by using word similarity among document index terms, and discusses the new model´s topology and information retrieval process. The experimental result shows that the extended model has better performance than the original one.