Two-way distributed sequential pattern mining using fruit fly algorithm along with Hadoop and map reduce framework

Abstract

Data mining is an effective tool used to take out information from big data as it provides several benefits to conquer the restrictions in it. In this paper, we present an innovative procedure developed using an updown directed acyclic graph (UDDAG) with fruit fly optimisation algorithm (FFOA), which is based on sequential pattern (SP) mining algorithm. In this work, the distributed sequential model mining algorithm is used to diminish the scanning time and scalability and the transferred database is employed to optimise the memory storage. The proposed method is used to expand the sequences in both the ends (prefixes and suffixes) of the identified model thereby supplying the consequences in quicker model expansion resulting in fewer database projections when compared to conventional methods. Our proposed method is implemented in Hadoop distributed surroundings to resolve the scalability issues and executed on JAVA platform using big datasets with Hadoop and map reduce framework.