The Research And Application Of Data Mining In Mining Rules Of Medical Diagnosis

With the rapid development of science and technology,various kinds of advanced medical testing equipment are constantly emerging, the informanization of hospital is continuously perfect and accumulated a lot of valuable medical information. However, facing with such huge volume of data, experts in medicine individually cannot analysis and summarize the knowledge in these data fully to support medical diagnosis. As a technology of knowledge acquisition , data mining resolve the difficult issue of knowledge acquisition with its own advantage, make full use of medical data accumulated by the medical platform. Thus, bringing data mining into computer aided medical diagnosis will greately promote the development of computer aided medical diagnosis and make it has a promising prospects.Firstly , introduced the data mining briefly and analied association rules carefully in this paper. In view of the most of actual medicial data for multiple-valued situation, given the improvement of association rules and proposed a algorithm of the quantitative association rules(Qarmasm). The algorithm is inspirited by the thought of discrimination matrix , proposed the concept of similar matrix; constructed similar matrix ,then get the per-frequent items and ID set that it belongs to through by traveling this matrix and generate association rules; so this algorithm can be mining data for multiple-valued situation without transition from multiple-valued data to bool data. we can generate per-frequent items and ID set that it belong to at the same time, so we only need operate incremental portion data and modify ID set that per-frequent items belong to while deleting or adding data form database, then generate association rules and proposed an effective incremental updata algorithm. So generated association rules without mining whole dataset, improvement improve efficiency and achieve a rapid, accurate incremental association mining algorithm. Then test the effectiveness of the algorithm from aspects of theoretical and experimental.At the end , constructed a platform of rules extraction used in computer aided diagnosis, and made a detail description about this platform from analysis, design to complementation.