Research Of Data Mining Based On Relational Rules

With the rapid development of computer science and technology, more and more people come to realize the value of information. So, people try to abstract useful information from a vast amount of data. And thus Data Mining comes in its way in such a situation. Comparing with the traditional statistic and query methods, Data Mining concerns multiple subjects, congregating some research results of artificial intelligence, pattern distinguishing, database, machine study and management information system, etc. Data mining is a newly-established frontier subject. It is being used extensively and its application future is bright.After studying and knowing of the basic character of Data Mining and some pertinent technology thoroughly, this paper analyzes some research results established and some flaws in Data Mining and carrys out some research about KDD and DM.Firstly, this paper introduces the situation of the development ofKDD and DM home and abroad, basic concepts and the construction and model of DM, then introduces the concept of relational rules and some algorithms of relational rules. Because of some flaws of those algorithms, a kind of Background-based Apriori algorithm is presented. In this algorithm we consider the bottleneck of getting frequent item-set and the interest of knowledge discovered and we add the fuzzy theory and the concept of semantics association rule to the algorithm. Then using the background knowledge, we convert the number property of database soundly. So knowledge dug out is more valuable and can be understood more easily. For the fuzzy transformation, the knowledge dug out is not limited in an accurate range of property only. And it has some useful applications. Finally, the algorithm is programmed to execute and the knowledge found can be visualized.