Study On Granular Computing Applying To Mining Association Rules In Database

In the recent years, with the development of information technology, the data we face become bigger and bigger. In order not to be puzzled by large quantity of data and get the knowledge we need, then, data mining was born. It's a newly-established frontier subject, congregating some research results of artificial intelligence, pattern recognition, database, machine learning and management information system, etc. It is being used extensively and its application future is bright.This thesis introduces association rules in data mining in details and analyse the hot spot in soft computing—granular computing. Granular computing as a new information and a method of dealing of knowledge was valued by many researcher and applied in many fields. This thesis applies granular computing technology to the area of mining association rules in data mining, studies the mining of association rules extensively from other ways. In the thesis, based on the analyzing and studying classic mining algorithms of association rules in details and summarizing its characteristic and limit through an instance, we introduce a model of mining association rules based on granular computing . All of the above rationally prepare for the proposition and construction of mining association rules algorithm based on the granular computing. Through defining information granule with binary string, we introduce an algorithm of mining association rules based on granular computing. Then we researched the methods of mining frequent items to get association rules using binary granules with two-dimension table and link structure. Based on the classic Apriori algorithm, we improve it and realize the algorithm of mining association rules based on the granular computing.In the end, the thesis compare the classic Apriori algorithm with the algorithm of mining association rules based on the granular computing through an experiment and analyze the result of it. The result proves the algorithm of mining association rules based on the granular computing is feasible and effective.