In order to solve the problem of low efficiency in traditional serial clustering integration algorithm in processing high-dimensional massive data

we propose a parallel clustering integration algorithm named SCEA (Spark based Clustering Ensemble Algorithm) which is based on spark platform. The input data of the SCEA algorithm is preprocessed by the combination of principal component analysis and pairwise constraints

which can reduce the dimension of the data and remove the feature correlation. After obtaining the base clustering results using different clustering algorithms

similarity matrix is constructed by the cluster labels of the base cluster members based on the triple method

and the hierarchical clustering algorithm is used to get the final clustering results. On the basis of calling the existing clustering algorithm in the spark MLlib

the SCEA algorithm is implemented based on Scala language. The SCEA is compared with other similar algorithms in multiple data sets. The experimental results show that SCEA is not only improved in accuracy than existing algorithms

but also proves the superiority of SCEA in algorithm performance by analyzing three performance indexes: running time

speedup ratio and scalability.