Application Of Data Mining In Medical Treatment Insurance

As the developping on the society and the computer techonology, it becomes nessary to change the management and operation of medical treatment insurance itself to more information. The medical treatment insurance management system in this paper is developed for the management and operation of the medical treatment insurance of Meihe City, and on the basis of the system we use the techonology of data mining in the medical treatment insurance field. Data mining ,that is,pick up the knowledge which people are interested in from the data of the great data-base. The knowledge picked up is connotative and unknown in advance,latent but useful,and expressed in the form of concept,rule,model etc. The knowledge can be supplied to decision-maker directly, used for assisting the course of the decision-making; or to expert in this field, used for modifing the knowledge system already existed of his own; or stored in the knowledge storing institution of application system as new knowledge. The course of data mining can be divided into three steps; a) Data preparation, include data integration, data choice and pre-analysis; b) Data mining, use synthetically data mining methods to analyse data in the data-base; c) Expression and explaination of the result, ananlyse information picked up on the base of the decision-aim of the final user, find out the most valuable information, and hand up the decision-maker. From different views, the data mining techonology has some methods of classifing: based on kinds of the knowledge picked up; on kinds of the data-base mined and on the techonology used. Classified based on kinds of the konwledge picked up. This kind of classify method includes: mining of summarize rules, mining of Characteristic Rules, mining of Association Rules, mining of Classification Rules, mining of clustering rules, trend analysis,difference analysis, modeling analysis etc..If divided by the abstract level, still has originality level data mining, high level data mining and mult-level ming and so on. And the usual used method of data mining are: Decision Tree method, Neural Network method, Genetic Algorithms, Statistic Analysis method, Rough Sets method, concept-tree method, Data Visualization and so on. At the beginning of data mining, it must prepare data, so we use the techonology of data warehouse. Now, the definition of data warehouse defined by the father of data-base W.H.Inmon in his book is:"Data Warehouse is a orient-thematic,compositive, steady,different-time data set, used for supporting the decision-making course in management."The four basic characters are: data in data warehouse is orient-thematic; data in data warehouse is compositive; data in data warehouse is steady; data in data warehouse change with time. Analysis and decision-making is on the base of data warehouse, and use the tool of OLAP. Based on the definition by OLAP committee, OLAP is a kind of software techonology, which made the analyser,manager or executer store data quickly,consistently and in turn. The data here is conformed from original data, which can reflect the dimension character of the enterprise and be comprehend by users. OLAP includes these basic concept: dimension, dimension level, dimension member, measurement, multi-dimension array. Basic analysis actions are: slice,dice, pivot,drill up,drill down.There are some request to OLAP in performace: quick, analysable,multi-dimension,information. Now in application, OLAP tool has two manners of organize data: one is OLAP application based on multi-dimension data-base(MOLAP), the other is OLAP application based on relation data-base(ROLAP). Due to the administrate system of multi-dimension data-bse is comparatively not mature and the widely application of relation data-base, ROLAP is the mainstream in the OLAP tools market. Some kinds of data organized form in data warehouse are: pile simply file, pivot synthesis file, simplify directly file and sequence file. The manner of data store in data warehouse can be divided into virual store,based on relation table and multi-dimension store. And the manner of based on relation table is used most common. Data warehouse design can't be divided clearly as data-base design, but can be divided in general into:concept model design, techonology prepartion, logical model design, physical model design, data warehouse builded,data warehouse run and maintenance.