Heterogeneous Multi core processors for improving the efficiency of Market basket analysis algorithm in data mining

摘要

Heterogeneous multi core processors can offer diverse computing capabilities.The efficiency of Market Basket Analysis Algorithm can be improved withheterogeneous multi core processors. Market basket analysis algorithm utilisesapriori algorithm and is one of the popular data mining algorithms which canutilise Map/Reduce framework to perform analysis. The algorithm generatesassociation rules based on transactional data and Map/Reduce motivates toredesign and convert the existing sequential algorithms for efficiency. Hadoopis the parallel programming platform built on Hadoop Distributed FileSystems(HDFS) for Map/Reduce computation that process data as (key, value)pairs. In Hadoop map/reduce, the sequential jobs are parallelised and the JobTracker assigns parallel tasks to the Task Tracker. Based on single threaded ormultithreaded parallel tasks in the task tracker, execution is carried out inthe appropriate cores. For this, a new scheduler called MB Scheduler can bedeveloped. Switching between the cores can be made static or dynamic. The useof heterogeneous multi core processors optimizes processing capabilities andpower requirements for a processor and improves the performance of the system.