In mixture distributions, mixing parameter is an important parameter. In this paper we estimate this parameter in mixture of two chi-square distributions using method of moments, EM-algorithm and Bayesian approach. These estimators are compared empirically; and their bias is investigated using numerical techniques. At the end, we conjecture that EM-algorithm is better than method of moments, but bayesian approach is better than method of moments and EM-algorithm. Analysis of a real data set is considered for illustrative purposes.
Digital Object Identifier (DOI)
Nasiri, Parviz and Azarian, AmirAbbas
"Estimating the Mixing Proportion of Mixture of two Chi-Square Distributions,"
Applied Mathematics & Information Sciences: Vol. 12
, Article 10.
Available at: https://dc.naturalspublishing.com/amis/vol12/iss4/10