关键词: 聚类, 密度峰值, 自适应密度峰值聚类, 相对局部密度, 相对连通距离

Abstract: ABSTRACT:Peak Clustering(DPC) was proposed in journal Science in 2014,which has aroused widespread discussion and application due to its efficiency and simplicity.However,some obvious shortcomings had been found in studies. To overcome these deficiencies,a novel clustering algorithm named Adaptive density peak clustering based on comparative quantities is proposed. In the improved algorithm, a new quantity named relative local density was used to assess the similarity between clusters,which greatly improved its applicability to datasets,another quantity called relative connectivity distance was applied for measuring the similarity between clusters,which effectively eliminates the influence of different sizes of clusters in the dataset. The applicability of the algorithm on different datasets was enhanced. By constructing a comentropy function, parameters could be adaptively determined according to the characteristics of the datasets,which improved the intelligence of the algorithm.A new allocation strategy is proposed to avoid the effects of the ‘chain reaction’.Simulations show that compared with the DPC and its improved algorithm,the performance of ACDPC algorithm is greatly improved.

Key words: clustering, density peak, adaptive density peak clustering, relative local density, relative connectivity distance