Normal Mixture Model-Based Clustering of Data Using Genetic Algorithm

44883

In this study, a new clustering algorithm was developed for the clustering of multivariate homogeneous and heterogeneous big data. Fragments in heterogeneous data determine the number and location of clusters. The number of fragments in heterogeneous data is determined based on both graphical and computational methods. In graphical methods, the probability graphs of each variable are used, while the computational methods use the univariate mixture normal distributions of each variable. Genetic algorithms are used to determine the location and structure of clustering centers corresponding to fragmentation in heterogeneous data. Determined models based on the number and structure of cluster centers is obtained by using mixture normal distributions. Each cluster center in mixture normal models corresponds to fragmentation in the variables. The best mixture model that matches the data structure from the mixture normal models is obtained by using the information criteria obtained from mixture normal distributions.

12

Genetic Algorithm , Gaussian Mixture Models , Model Based Clustering , Information Criteria

Erciyes University Journal Of The Institute Of Science an‎d Technology

23

Erciyes University Journal Of The Institute Of Science an‎d Technology