A New DBSCAN Parameters Determination Method Based on Improved MVO

Density-based spatial clustering of applications with noise (DBSCAN) is a typical kind of algorithm based on density clustering in unsupervised learning. It can cluster data of arbitrary shape and also identify noise samples in the dataset. However, an unavoidable defect of the DBSCAN algorithm exists since the clustering performance is quite sensitive to the parameter settings of MinPts and Eps, and there is no theory to guide the setting of its parameters. Therefore, a new method is proposed to optimize the DBSCAN parameters in this paper. Multi-verse optimizer algorithm, a special variable updating method with excellent optimization performance, is selected and improved for optimizing the parameters of DBSCAN, which not only can quickly find out the highest clustering accuracy of DBSCAN, but also find the interval of Eps corresponding to the highest accuracy. In order to search the range of Eps more quickly and efficiently, we design a new mechanism for the variable update of MVO. The experimental results show that the improved MVO is used to optimize DBSCAN, which not only can quickly find out its highest clustering accuracy but also can search the parameters of MinPts and Eps corresponding to the highest clustering accuracy efficiently.

Abstract licence: CC BY-NC-ND

33 more references, whose wording the paper’s licence keeps private.

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.