A privacy-preserving group decision making expert system for medical diagnosis based on dynamic knowledge base

Abstract

Most of the present researches on wise medical incorporate a static supporting data, while medical knowledge and rules are featured with a dynamic change, which affect calculation efficiency and accuracy of the result. This paper combined data mining technology with expert system framework, established a medical expert system with a dynamic knowledge base. In the innovative dynamic system, samples were together with their classification results would be added to the knowledge base. In addition, privacy protection was considered in the expert system by means of data removal and conversion of attribute value. At last, for a comprehensive and precise evaluation of both numeric and character attribute data, we proposed a new similarity measurement algorithm termed as E–JD, which combinate Euclidean Distance algorithm with Jaccard Distance algorithm. For more improvement of the data classifasion accuracy, based on KNN and the dynamic knowledge base, the innovative algorithm of D-SNN was proposed, with a modification in the selection and decision of the nearest neighbor objects. The experimental comparison between 1NN, KNN, KNN-means and D-SNN algorithm, shows that the classification performance of the proposed algorithm, D-SNN, exhibits an impressive improvement.