Medical Record Retrieval Based On Knowledge Of Medicine

With the extensive application of computer technology and digital imaging technology in the medical domain, medical case has become digitalized and multi-modal. Meanwhile, the medical data generated by hospitals and medical center grows rapidly, but these data has not been effectively used. From the perspective of computer-aided diagnosis, the efficiency and accuracy of diagnosis will be improved if the cases similar to the current case can be retrieved from the case database. Therefore, medical case retrieval which has a great research value and application prospective, has become an important research topic in computer-aided diagnosis, and draw more and more researchers’attention.This thesis studies text-based medical case retrieval, and explores knowledge-based methods for medical case retrieval. After reviewing and analyzing the previews works, this thesis introduces the basic theories of information retrieval and the commonly used ontologies and related tools, then it compare several retrieval models and document represents to find the suitable ones for medical case retrieval, finally,it proposes two retrieval methods,which using of medical ontologies for query expansion and document expansion. Both the two proposed methods can significantly improve the performance of medical case retrieval.The main contributions of this thesis include:Firstly, explore and compare various document representations and retrieval models for medical case retrieval to find the optimal configuration.Secondly, expand query and document with the MeSH terms to improve the retrieval performance of medical case retrieval.Thirdly, use phrase-based retrieval model and expand query with UMLS to improve the performance of retrieval.