Abstract: In clustering analysis,one of the important problems is mixed data clustering.The clustering of existing algorithms is mainly based on similarity measurement of all samples.Therefore,the efficiency of clustering for large-scale data is not high.So we designed a new sampling strategy and proposed an ensemble algorithm for large-scale mixed data based on sampling.This new algorithm clusters subsets which are obtained by the use of the new sampling strategy respectively and the final clustering results can be gotten by clustering ensemble.Experiment shows that the efficiency of algorithm is improved significantly and the clustering validity indexes are almost the same compared with the modified K-prototypes algorithm.

Key words: Clustering,Large-scale mixed data,Clustering ensembles,Sampling,Validity index