Frequent Data Mining in Data Publishing for Privacy Preservation

Abstract

Weighted frequent pattern mining is suggested to find out more important frequent pattern by considering different weights of each item. Weighted Frequent Patterns are generated in weight ascending and frequency descending order by using prefix tree structure. These generated weighted frequent patterns are applied to maximal frequent item set mining algorithm. Maximal frequent pattern mining can reduces the number of frequent patterns and keep sufficient result information. In this paper, we proposed an efficient algorithm to mine maximal weighted frequent pattern mining over data streams. A new efficient data structure i.e. prefix tree and conditional tree structure is used to dynamically maintain the information of transactions. Here, three information mining strategies (i.e. Incremental, Interactive and Maximal) are presented. The detail of the algorithms is also discussed. Our study has submitted an application to the Electronic shop Market Basket Analysis. Experimental studies are performed to evaluate the good effectiveness of our algorithm..