Applications Of Machine Learning In The Acquisition Of Medical Insurance Decision-making Knowledge

Since 1998, the Medical Insurace System has been established in China. The Medical Insurance Agencies have been handling to explore the solutions of the contradiction of the medical insurance fund pressure and the increasing medical needs. The decision-making knowledge of the Medical Insurance parameters adjustment comes from the careful analysis of real business data. It becomes the new direction how to extract accurate statistical analysis, forecasting from the heterogeneous data.This article describes the knowledge of Medical Insurance decision-support from a knowledge engineer's viewpoint, and tries to decompose the complex process of acquisition of Medical Insurance decision-making knowledge into a series of individual steps. For the key steps, we use machine-learning technology to learn the knowledge automatically from the business data of over the years as training data. According to different business needs, analyses the different process of getting different expertise, and the appropriate machine learning techniques. Realize the automation of the Medical Insurance Decision-making Knowledge engineering.As examples, in this article, we research how to get the knowledge of the evaluations of fund balance risks of enterprises and the clinic structures in detail. We construct the learning algorithm based on decision tree of the acquisition of the evaluations of fund balance risks of enterprises, and discuss the questions of over-matching, properties-discretion and properties-contribution. We construct six models of clinic structures; they are hospital-cycle models, cost models, check modes, operation modes, medicine models and special materials models, and the submodels such as the total cost models, bed-day cost models, cost distribution models, commonly used medicine models, medicine compatibility models. For these models, we mine the common-item models and the associations by using frequent intemsets mining algorithm, and learn the knowledge of the differences of clinic structures by using LMS algorithm, and approach the accuracy and the efficiency.