A Parallel DBSCAN Algorithm Based on Spark

摘要

With the explosive growth of data, we have entered the era of big data. In order to sift through masses of information, many data mining algorithms using parallelization are being implemented. Cluster analysis occupies a pivotal position in data mining, and the DBSCAN algorithm is one of the most widely used algorithms for clustering. However, when the existing parallel DBSCAN algorithms create data partitions, the original database is usually divided into several disjoint partitions, with the increase in data dimension, the splitting and consolidation of high-dimensional space will consume a lot of time. To solve the problem, this paper proposes a parallel DBSCAN algorithm (S_DBSCAN) based on Spark, which can quickly realize the partition of the original data and the combination of the clustering results. It is divided into the following steps: 1) partitioning the raw data based on a random sample, 2) computing local DBSCAN algorithms in parallel, 3) merging the data partitions based on the centroid. Compared with the traditional DBSCAN algorithm, the experimental result shows the proposed S_DBSCAN algorithm provides better operating efficiency and scalability.