A Study on Data Mining Theoretical Frameworks Based on Correlativity Sets
基于相关集合的数据挖掘理论基础研究

The plausibility relation which is generalization of fuzzy relation and probabilistic relation is proposed in the paper. We think data mining to be a process of finding the plausibility relation in database and correlativity measure to be a particular plausibility relation based on correlativity sets. The critical calculates such as the accuracy of the rough sets, the confidence and the bayesian form in data mining can be united using the correlativity measure. The GPDM (General Process of Data Mining)that represents the nature of data mining is also proposed. The data mining theoretical foundation and frameworks based on correlativity sets are also given and discussed in the paper.