Incremental general regression with expectation maximization for learning finite mixtures using data with missing values
Ahmed R. Abas · 2013
Finite mixture models (FMM) is a pattern recognition method, in which parameters are determined from complete data using the Expectation Maximization (EM) algorithm. This paper presents an algorithm for determining parameters of the finite mixture models using data having missing values. Compared with a developed EM algorithm that is proposed earlier the proposed algorithm has proved good performance when the features containing missing values are at least moderately correlated with some of the complete features in the input data set.