MINING THE INCREMENT ASSOCIATION RULES IN TEMPORAL DATABASE

Temporal Database is one of the important data information systems. There is some useful unknown knowledge concering the tendency and the relation of data increment. In this paper the definitions of adjacency, increment, collision set are given and the model of discretization ＆ code based increment association rules are proposed. Some discretization methods for quantitative attributes are compared, which includes partial completeness, entropy based discretization, C4.5, and rough sets based full discretization. Finally, the algorithm TIDM that joins an ameliorative method RBWAD of rough sets based full discretization with general data mining is discussed.