The density peak clustering（DPC）algorithm can quickly search and cluster density peaks

it has the characteristics of less adjustment parameters and no iteration.But for the data of sparse clusters and dense clusters

simple local density definition cannot find the cluster center effectively

and property of fault tolerance is poor.In order to solve the above problems

a density peak clustering algorithm based on micro-cluster fusion（MCF-DPC）is proposed.Firstly

the local density calculation formula is improved.Secondly

the micro-cluster fusion strategy is combined with DPC.Finally

a non-negative decomposition algorithm is introduced to reduce the dimensionality of the data.Experiments on multiple datasets verify the correctness and effectiveness of the improved algorithm.