A New Model for Multiparty Collaborative Distributed Mining using Bloom Filter

Collaborative pattern mining is a pattern mining technique used in a distributed environment where each site participate collaboratively with one another for the discovery of patterns. Each individual site is capable of mining patterns and can communicate with other sites in the distributed environment. Currently existing work have failed to tackle various problems using this technique like effective data exchange and cross-site pattern pruning mechanisms. This paper advocates an effective method to perform pattern pruning in a distributive manner with the help of a data structure called Bloom Filter. With its help cross site pattern pruning can be performed with minimized data transfer between the sites as well as increased speed and with the capability of mining Local, Global and Inter-database Patterns. Keywords— Distributed data mining, Bloom Filter, Frequent item set, Distributed associative rule mining