Interactive Mining of Maximal Constrained Frequent Patterns

Frequent pattern mining is an important part in data mining. Recently, in order to improve the speed of mining, a lot of constraint-based algorithms are presented. But the study of interactive mining is not sufficient. In this paper, a novel algorithm called interactive mining of maximal constrained frequent patterns (IMCP) is presented. The FP_tree is constructed according to the descending or ascending order of constraints (In this paper, the constraints are price and quantity) with scanning database once. In addition, it allows users to dynamically change constraints during the process. Experiment proved that IMCP is efficient and scalable.