Adaptive density peak clustering algorithm combined with sparse search

Abstract With the advantages of few parameters and ability to deal with clusters arbitrary shape, density peak clustering algorithm has attracted wide attention since it came out. However, problems such as high time complexity, poor effect on complex data sets, need manually select cluster centers. Aiming at above shortcomings, an improved is proposed. Combined sparse search algorithm, calculation similarity between each point its nearest neighbor simplified, problem complexity overcome. A new local definition method adopted make points better reflect spatial structure distribution improve accuracy algorithm. Finally, a strategy for automatically selecting centers proposed adaptability The used compare other artificial sets real sets. experimental results show that can quickly accurately identify various clusters.