Research On Association Rules Mining Based On Temporal Data

Association rule is a main model of the data mining. It has become an very important research topic in data mining field. Association rule mining aims to confirm the relations of different attribution in masses of data, furthermore, to find out the relying relations between various attributions, which can meet the need of the support threshold value and reliability threshold value.The reality world is a variable world. Time, as a basic part, can reflect the information of the reality world. So most datas have temporal attributions. Nowdays more and more people are interested in the temporal data mining. The temporal association rule mining has become the issue of many researchers.First, we discuss the definition, technology, and describe the researching background and researching content in this paper.Time is an important attribution of the temporal datas. It is helpful to define an scientific definition to express time in temporal database and found temporal model in temporal data mining. So we introduce the basic concepts about temporal type, temporal factor and temporal granularity. So we propose a kind of mathematical model about temporal association rules in transaction database, and describe several temporal association rules with practice significance.Finally, based on the research of indeterminate time, and introduce a expressing method of indeterminate time, we propose simultaneous temporal relations of indeterminate time and intersecting of indeterminate time. Following the case of temporal data example, we make a new mining method of indeterminate temporal data of a kind of valid time and the intersecting of indeterminate time, and we propose a corresponding arithmetic.As a result, the relations between cases are attained.