Research On Mining Technology Of Association Rules And Meta-Rules

The rapid advancement of the computer technology has driven the applications of computer information management system in every walk of life. A large amount of data has been accumulated in databases. These data are like a gold mine of knowledge, which has proved invaluable in helping various decision-making processes. So how to discover the knowledge and how to discover the knowledge effectively became an important research area in computer science. The researchers in this area began to study the Knowledge Discovery in Database,that is KDD, in the late 1980's. The First International Conference of Knowledge Discovery and Data Mining at which the field of Data Mining was first put forward was held in Montreal in Canada in 1995.Mining association rules is an important task of data mining, which is also the focus of this article. Existing research work for association rules mining focuses mainly on the efficiency of the mining methods, whereas the quality of the mined association rules has received little attention. An example of the importance of data mining quality is time-series databases. In these databases, the intensity of the association rules may change with time. This leads to the problem that the association rules derived from existing algorithms may not be applicable in future situations. Therefore, the application of association rules can be difficult. On the other hand, existing association rule mining algorithms cannot identify the trends of association rules so that some useful rules might be overlooked. Aiming at solving these problems of the existing association rule mining algorithms, this paper discusses the basic idea and the method to mine association rules and meta-rules synthetically. An approach to formalize the meta-rules of association rules is presented. An integrated algorithm for mining association rules and meta-rules is proposed. It is also discussed how to classify the meta-rules based on machine learning strategies.We designed an algorithm for mining frequent patterns using multi-hash...