Incremental CFP-Tree Optimization For Efficient Representative Pattern Set Mining

In the area of data mining frequent pattern mining is a significant hitch. Frequent pattern mining is more often performed on a transaction database that contains set of items. A pattern is frequent pattern when it has bigger support than user define threshold. For mining frequent patterns numerous capable algorithms have been developed. However, RPglobal is awfully time-consuming and space-consuming. Barely it becomes realistic when the number of frequent patterns is not large. RPlocal is dreadfully efficient, although it produces extra representative patterns than RPglobal. Here, two algorithms MinRPset and FlexRPset are in picture. Algorithm MinRPset is comparable to RPglobal, but it utilizes numerous techniques to diminish the running time and memory usage. CFP-tree structure is used in MinRPset, it is a tree structure. FlexRPset provides one parameter K in addition, which allows users to build a swap between efficiency and the number of representative patterns that he has chosen. Use of the techniques is supportive to improve the effectiveness of MinRPset by considering closed patterns only and by using a structure called CFP-tree to find C (X) s efficiently.