sample nature of the change point estimator and the mixture parameter estimator. In order to verify theР effectiveness of the method, we conducted some simulation experiments. The results show that the EMР algorithm that does not consider the change point has a poor estimate on the classification result, andР even loses some categories; Our improved EM algorithm can accurately locate the location of the changeР point, and at the same time obtain accurate estimates for the corresponding parameters of each category.Р Keywords: change point; mixture model; multi-classification; EM algorithmР 2020 Mathematics Subject Classi?cation: 62-08