Interest Incremental Data Mining Algorithm

Data mining is a kind of technique which gets data from database and makes intelligent fact-based operation decision and strategy by using advanced analysis method,it's an effective method for getting information from mass data. Temporal data is an important type of information. This paper studies on increment data mining of temporal data and introduces the concept of interest measure as its evaluating standard. This paper falls into the following three parts.In the first part,we expound the present situation of data mining,concrete mining process and the classification of data mining method,figure out concrete flow chart,explain the specific model of association rules,and present how Apriori arithmetic mined the Boolean association rules of transaction.In the second part,we introduce interest measures and discuss the interests of association rules,which are mainly evaluated from the subjective and objective points.We distinguish the rules which interest the users from those which disinterest the users by templates. By this we accomplish the subjective evaluation of the interest of association rules. Later,we put forward the concept of interest measure of association rules on the basis of the confidence and support number,By this means,constraints are added to objective evaluations of the interest of association rules.In the last part,we give the meaning of adjacency and increment in temporal database,put forward the basic increment model which based on discretization,introduce several methods for discretization,and constitute an integrated data mining model on the basis of combining with DHA arithmetic. Finally the quantitative association rules which interest the users by making use of some standards such as interest measure are elicited.The correlative models,algorithms and analysis have been proved to be effective and feasible,and its resultant operation models and rules are able to provide effecient information for systemic decision-making.