Local density based on weighted K-nearest neighbors for density peaks clustering

Sifan Ding, Min Li, Tianyi Huang, William Zhu

Abstract: Highlights•A new similarity measure accurately identifies the nearest neighbor for each point.•Each point is assigned a weight, representing its probability of being a K-nearest neighbor to other points.•Local density is calculated using weighted K-nearest neighbors.•The new method enables the detection of density peaks in regions with arbitrary complexity.