Mixed attribute clustering algorithm based on filtering mechanism

For the existing mixed data clustering algorithm, there are some problems such as low clustering accuracy and parameters sensitive, this paper proposes a mixed attributes data clustering algorithm (MC-FM) based on filtering mechanism. The algorithm measures the similarity between objects with improved mixed attribute similarity, and uses the local mean shift of each object, which based on KNN and mean shift, then distinguishes the core object and the non-core object according to the filtering mechanism. Finally, the non-core objects are divided into corresponding clusters to form the final cluster. The experimental results of the algorithm on the synthetic dataset and UCI dataset verify the effectiveness of the algorithm. Compared with similar algorithms, MC-FM has higher clustering accuracy.