DBSCAN Clustering Algorithm Based on Density

Abstract

Clustering technology has important applications in data mining, pattern recognition, machine learning and other fields. However, with the explosive growth of data, traditional clustering algorithm is more and more difficult to meet the needs of big data analysis. How to improve the traditional clustering algorithm and ensure the quality and efficiency of clustering under the background of big data has become an important research topic of artificial intelligence and big data processing. The density-based clustering algorithm can cluster arbitrarily shaped data sets in the case of unknown data distribution. DBSCAN is a classical density-based clustering algorithm, which is widely used for data clustering analysis due to its simple and efficient characteristics. The purpose of this paper is to study DBSCAN clustering algorithm based on density. This paper first introduces the concept of DBSCAN algorithm, and then carries out performance tests on DBSCAN algorithm in three different data sets. By analyzing the experimental results, it can be concluded that DBSCAN algorithm has higher homogeneity and diversity when it performs personalized clustering on data sets of non-uniform density with broad values and gradually sparse forwards. When the DBSCAN algorithm's neighborhood distance eps is 1000, 26 classes are generated after clustering.