Abstract:

In the process of diagnosis and treatment service for different types of diverse patients， how to achieve rapid intelligent diagnosis according to the patient’s condition is a new challenge and frontier problem in medical services. Based on this， the complicated and varied diagnosis information and diagnosis process of patients are analyzed， then an intelligent diagnosis and treatment decision method that considering the multi-source behavior fusion information generated by patients is proposed in this paper. In the proposed intelligent diagnosis decision method， the attribute diversity of different types of patient diagnosis and treatment information is considered， and a distance measurement method considering different types of attribute features is defined. Then， on the basis， a rough set attribute weight determination algorithm based on similarity measure is used to calculate the weight of each attribute. Further， the similarity measurement and consistency test of attributes are combined to compare and analyze the information of target patients and historical patients. According to the similarity value obtained by quantitative calculation in the reasoning process， the most similar case to the target patients can be obtained， which can help doctor give accurate treatment plans. Finally， the validity and feasibility of the proposed method are verified by UCI heart disease data from Cleveland data set.

Key words: case-based reasoning, multi-source fusion behavior information, rough set, similarity measurement, consistency test