Divides according to the data mining duty, has the following several kinds: The classification or the forecast model data mining, the data summarize, the data gathers kind, the connection rule discovery, the sequence pattern discovery, the dependent relations or the dependent model discovery, exceptionally with tendency discovery and so on.
The connection rule excavation is in a data mining important constituent. It mainly is seeks assigns the data set 中項 between the interesting connection and the correlation relation. The most famous connection rule discovered the method is the Apriori algorithm which R.Agrawal proposed. The connection rule discovery may divide into two step. First step is iterates the recognition all frequent projects collection, the request frequent project collection support rate is not lower than the user hypothesis the minimum value; Second step is not lower than the user hypothesis from the frequent project centralism structure confidence level the minimum value rule. The recognition or the discovery all frequent projects collection is connected the rule discovery algorithm the core, also calculates the quantity most major part.
This article mainly conducts the research on the connection rule analysis, including elementary theory and algorithmic analysis. The main work includes: The analysis connection rule classical algorithm, points out its good and bad points and the applicable scope; To classical algorithm: The Apriori algorithm programs the realization with C; Carries on the experiment with the many kinds of data sets to the algorithm, and has analyzed the experimental result. Finally above the algorithm will integrate in together, the compilation visible contact surface, has realized a small data mining system.
Key word: Data mining; Gathers a kind of analysis; Connection analysis; Algorithm