基于独立的Gaussian与Beta有限维混合模型的聚类算法

Joint Finite Mixture Modelfrom Independent Gaussian and Beta Distributed Data

对基于独立的Gaussian与Beta有限维混合模型的聚类算法进行了探讨与研究,并在此基础上提出一种新的聚类算法-BGMMn聚类算法,给出算法流程。通过数据模拟比较了四种聚类算法(BGMMn,BGMMs,BGMMa,BGMMh)对聚类数目估计的准确度,BGMMn算法的准确度优于其他3种聚类算法。

The article has carried out the joint finite mixture model from independent Gaussian and beta distributed data, I develop a new clustering algorithm, that called BGMMn, and give up the algorithm processes. By the simulation, giving the comparison between four clustering algorithms （BGMMn, BGMMs, BGMMa and BGMMh）on the correct number of clusters, BGMMn algorithm accuracy better than other three clustering algorithm.