A New Density Peak Clustering Algorithm With Adaptive Clustering Center Based on Differential Privacy

A new density peak clustering (DPC) algorithm with adaptive center based on differential privacy was proposed to solve the problems of poor adaptability high-dimensional data, inability automatically determine centers, and in analysis. First, problem cosine distance used measure similarity between datasets. Then, aiming at subjective selection, from perspective ranking graph, weight $(i-1)/i$ introduced creatively, slope trend graph redefined realize center. Finally, problem, Laplacian noise appropriate budget added core statistic (local density) achieve balance protection effectiveness. Experimental results both synthetic UCI datasets show that this can not only automatic selection center, but also analysis, improve evaluation index greatly, which proves effectiveness algorithm.