Robust fitting of mixture models using weighted complete estimating equations

Mixture modeling, which considers the potential heterogeneity in data, is widely adopted for classification and clustering problems. models can be estimated using Expectation-Maximization algorithm, works with complete estimating equations conditioned by latent membership variables of cluster assignment based on hierarchical expression mixture models. However, when components have light tails such as a normal distribution, model sensitive to outliers. This study proposes method weighted (WCE) robust fitting Our WCE introduces weights that automatically downweight The are constructed similarly density power divergence models, but our WCE, they depend only component distributions not whole mixture. A novel expectation-estimating-equation (EEE) algorithm also developed solve WCE. For illustrative purposes, multivariate Gaussian mixture, experts, skew considered, how EEE implemented these specific described. numerical performance proposed estimation was examined simulated real datasets.