earticle

영어

Mining association rules for data is not only an essential part of data mining but also a hot issue in knowledge engineering and researches on data mining technology. Since multi-data mining is characterized as being multi-type, multi-level, multi-implicational and complicated, the efficiency of multi- data association rule mining usually cannot be high, precision and accuracy are of a relatively low degree, and the targets of mining cannot be obtained quickly. Therefore, on the basis of improving traditional association rule mining algorithm, this paper researched on multi-data association rule mining algorithm and based on grey relational analysis, proposed a multi-data association rule mining algorithm. Firstly, the associate objects most relevant to the target objects are obtained through the grey relational analysis, which helps to form single- or multi-target data associates; after that, the multi-data association rule mining model which sets data associates as the new mining objects is established. Under the conditions that the level of support and confidence are met, the frequent patterns of corresponding data associates and further, the multi-data association rule, are obtained. Simulation experiments implied that the model have the advantages of simplicity, practicality, operability, decent precision and accuracy.