Preserving Privacy of Association Rule Mining in Horizontally Partition Database

Data mining is very important and most growing area today and that is used to find or extract important knowledge from large data collection. These data collection are partitioned among various sites or parties. Privacy may avoid the parties from directly sharing the data. So, we use a protocol for preserving privacy of association rule mining in horizontally partition database. The current leading protocol is Kantarcioglu and Clifton as known a K & C protocol. This is based on unsecured distributed version of the apriori algorithm named as Fast Distributed Mining (FDM) algorithm of Cheung et al. The main part of this protocol has two novel secure multi- party algorithm one that computes the union of private subset that is each of interacting player hold and another that tests whether an element held by one player is included in a subset held by another. In that setting there are various sites or player that hold homogeneous database, means that share the same schema but hold the information on different entities. The main goal of this is to find out frequent itemsets and association rule with minimum support threshold and minimum confidence threshold. Data mining association rule technique is used for discovering or finding interesting relations in large database. Support is how frequently a specific item or itemsets are present in database and confidence is an indication of how often the rule has been found to be true. This protocol offers enhanced privacy with respect to early protocol. In addition it is not complicated and it is very vary effective in the terms of communication cost, communication round and last computational cost.