Computer Science · Artificial Intelligence
Density Peaks Clustering Algorithm Based on Weighted <i>k</i>-Nearest Neighbors and Geodesic Distance
See on the mapAbstract
Density peaks clustering (DPC) is a density-based clustering algorithm with excellent clustering performance including accuracy, automatically detecting the number of clusters, and identifying center points. However, DPC has some shortcomings to be addressed before it can be widely applied. For example, sensitive predefined parameter is not suitable for manifold datasets, and decision graph easily causes the wrong center points. To address these issues, a new DPC algorithm based on weighted k-nearest neighbors and geodesic distance (DPC-WKNN-GD) is proposed in this article to improve the clustering performance for manifold and non-manifold datasets. The DPC-WKNN-GD introduces the weighted k-nearest neighbors based on Euclidean distance to optimize the local density <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\rho $ </tex-math></inline-formula> . Inspired by DPC-GD, the DPC-WKNN-GD redefines the distance <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\delta $ </tex-math></inline-formula> based on geodesic distance. The experimental results on artificial and real-world datasets, including image datasets, show that the DPC-WKNN-GD outperforms the state-of-the-art comparison algorithms on both manifold and non-manifold datasets. In addition, the DPC-WKNN-GD completely overcomes the DPC-GD defect, in which the decision graph displays misleading center points.