基于MapReduce计算模型的并行关联规则挖掘算法研究综述

关联规则挖掘是最常用、最重要的数据挖掘任务之一,经典的关联规则挖掘算法有Apriori、FP-Growth、Eclat等。随着数据的爆炸式增长,传统的算法已不能适应大数据挖掘的需要,需要分布式、并行的关联规则挖掘算法来解决上述问题。MapReduce是一种流行的分布式并行计算模型,因其使用简单、伸缩性好、自动负载均衡和自动容错等优点,得到了广泛的应用。本文对已有的基于MapReduce计算模型的并行关联规则挖掘算法进行了分类和综述,对其各自的优缺点和适用范围进行了总结,并对下一步的研究进行了展望。

Parallel Association Rules Mining Algorithm Based on MapReduce : A Survey

Association rule mining is one of the most commonly used, important task of data mining, classic algorithm for mining association rules are Apriori, FP - Growth, Eclat, etc. With the explosive growth of data, traditional algorithms cannot meet the needs of the large data mining, Distributed, parallel algorithm for mining association rules are needed to solve the problem of mining association rules in large data.MapReduce is a kind of popular distributed parallel computing model, Because of its simple to use, good scalability, the advantages of automatic load balancing and fault tolerance, has been widely used. In this paper, the existing parallel algorithm for association rules minging based on mapreduce are classified and reviewed, to their respective advantages and disadvantages and scope of application are summarized, and the next research is prospected.