Research On Data Mining Based Decision Rules And Association Rules

Nowadays, stored data is increasing violently. However, it is well known that acquiring knowledge is a bottleneck. Therefore, how to acquire related knowledge from "Data Ocean"efficiently has become a core issue. With the advent data mining techniques, precise knowledge has been detected effectively. After the architecture, concept, research status and potential application foreground of Data Mining techniques are summarized and the basic ideal related to this paper are expatiated; discussing the problems of mining quantificational association rules, discussing the problems of decision rules based on the ideal of Rough set theory and developing a prototype system. The contents of the paper are listed as follow: On the field of quantificational association rules: Firstly, on the basis of the new mapped Boolean database, quantificational association rules can be acquired. Secondly, the model of association rules and classical Apriori algorithm are introduced. In order to improve its efficiency, an algorithm called HBA is proposed which introduces the Hash tree. On the field of decision rules based on the ideal of Rough set theory: Firstly, on the basis of comprehending complement algorithm of ROUSTIDA, it is applied to realize the complement of an incomplete information table. Secondly, an inductive learning approach based on modified rough set is proposed. On the one hand, the continuous attributes in the decision table are fuzzified with the proper fuzzy membership functions and the fuzzy similar matrix of the attributes is constructed with the fuzzy degree of nearness, then k-w method is applied to evaluate the relative importance of every continuous attribute. The continuous decision table is discretized into a compatible table based on the fuzzy similarity relation. On the other hand, an improved definition of the attribute significance based on the weighed sum is proposed. Based on the key technologies stated above, a prototype system is developed. Several illustrative examples have demonstrated the effectiveness and feasibility of the proposed methodologies.