Combining concept learning and probabilistic information retrieval model to understand user’s searching intent in OWL knowledge base

Abstract

Understanding and describing user’s searching intent in exploratory information retrieval is a key issue for improving the relevance of search results. Employing concept learning method and probabilistic information retrieval model, this paper proposes an exploratory information retrieval strategy that can explain user’s search intent in a formal way. User’s relevance feedback from the initial search results are considered as examples and the user’s searching intent is described as concepts learned from the knowledge base and examples. Uncertain inference with respect to the concept learned in knowledge base is used to implement probabilistic information retrieval. By constructing a probabilistic OWL knowledge base, this paper develops a healthcare interactive information retrieval prototype to evaluate the method proposed. The experiment results prove the advantages of using concept learning in exploratory semantic retrieval.