Mining association rules is an important data mining problem which is the interested, hidden and unknown knowledge that can be discovered from large data. After analyzing and studying association rules mining algorithms based on support-confidence framework, we found that there are two problems existing in them. 1. It is probable that the rule will possess high support and confidence, but ~vill be unintercsting, even be false. 2. It can't produce rules with negative items. To solve the two problems, the paper first introduces the third threshold value to increase interest measure ----correlation support. It is regarded as an effective pattern when produced support, confidence and correlation support of an association rule is simultaneously greater than minimum support, confidence and correlation support. Second, in order to produce rules with negative items is added negative itemscts in which correlation support is less than one. A good and effective association rules mining algorithm is ~vorked out by the above improvement. It realized only providing the interested pattern and improved associative mining. |