Proportional data modeling via selection and estimation of a finite mixture of scaled Dirichlet distributions. (19th November 2019)

• Record Type: Journal Article
• Authors: Zamzami, Nuha; Alsuroji, Rua; Eromonsele, Oboh; Bouguila, Nizar
• Is Part Of: Computational intelligence. Volume 36:Number 2(2020)
• Page Start: 459
• Page End: 485
• Publication Date: 2019-11-19
• Subjects: mixture modeling -- model selection -- proportional data clustering -- scaled Dirichlet distribution -- unsupervised learning; Artificial intelligence -- Periodicals; Computational linguistics -- Periodicals; 006.3
• DOI: 10.1111/coin.12246
• Languages: English
• ISSNs: 0824-7935

Abstract

Abstract: This paper proposes an unsupervised algorithm for learning a finite mixture of scaled Dirichlet distributions. Parameters estimation is based on the maximum likelihood approach, and the minimum message length (MML) criterion is proposed for selecting the optimal number of components. This research work is motivated by the flexibility issues of the Dirichlet distribution, the widely used model for multivariate proportional data, which has prompted a number of scholars to search for generalizations of the Dirichlet. By introducing the extra parameters of the scaled Dirichlet, several useful statistical models could be obtained. Experimental results are presented using both synthetic and real datasets. Moreover, challenging real‐world applications are empirically investigated to evaluate the efficiency of our proposed statistical framework.