A0396
Title: Finite mixture model based on the GSMMGN family with several interval censoring
Authors: Ruijie Guan - Beijing University of Technology (China) [presenting]
Tsung-I Lin - National Chung Hsing University (Taiwan)
Weihu Cheng - Beijing University of Technology (China)
Abstract: The generalized scale mixtures of mixture generalized normal (GSMMGN) distribution is presented, which is a versatile family of distributions capable of modeling data with diverse and flexible shapes. A novel finite mixture model based on the GSMMGN class of distributions with several interval censoring (FM-GSMMGN-SIC) is established, which provides a basic framework for modeling complex data exhibiting multimodality, large skewness, heavy tails, leptokurtic or platykurtic behaviors, and missing values simultaneously. A variant of the EM-type algorithm is formulated by combining the reparameterization technique and profile likelihood approach (PLA) with the classical Expectation Conditional Maximization (ECM) algorithm for parameter estimation of the proposed model. This approach with analytical expressions in the E-step and tractable M-step can greatly enhance the computational speed and efficiency of the algorithm. Some simulation studies are conducted to assess the performance of the proposed algorithm, and the results show satisfactory outcomes for several artificial datasets. Moreover, the feasibility and practical usefulness of the proposed methodology are illustrated through real data analysis.
