A Parallel Clustering Method Study Based on MapReduce

摘要

Clustering is considered as one of the most important tasks in data mining. The goal of clustering is to determine the intrinsic grouping in a set of unlabeled data. It has been widely applied to many kinds of areas. Many clustering methods have been studied, such as k-means, Fisher clustering method, Kohonen neural network and so on. In many kinds of areas, the scale of data set becomes larger and larger. Classical clustering methods are out of reach in practice in face of big data. The study of clustering methods based on large scale data is considered as an important task. MapReduce is taken as the most efficient model to deal with data intensive problems. In this paper, parallel clustering method based on MapReduce is studied. The research mainly contributes the following aspects. Firstly, it determines the initial center objectively. Secondly, information loss is taken as the distance metric between two samples. The efficiency of the method is illustrated with a practical DNA clustering problem.