| Summary: | 碩士 === 慈濟大學 === 醫學資訊學系碩士班 === 103 === A huge amount of biomedical evidences about genes and diseases have been published in literature. To build and maintain online databases of the evidences about gene-disease associations, expert curators often strive to find out the reference about gene-disease associations from the ever-increasing huge number of biomedical references. In this thesis, we present a technique called CRFT (ranking conclusive, rich and focused texts) to rank the biomedical references so that those references that are really related to gene-disease associations can be ranked high for the curators to read and check. CRFT ranks the references by integrating three measures: degree of conclusiveness, degree of richness, and degree of focus. We also evaluate CRFT using more than one hundred thousand references for over one thousand gene-disease pairs. CRFT performs significantly better than other techniques in ranking references. It can thus support biomedical experts in curating gene-disease associations in a timely complete manner.
Keywords: Gene-disease association, Reference ranking, Conclusiveness of information, Richness of information, Focus of information
|
|---|
