BAGGING AND SUBAGGING IN MIXTURE MODELS

Two bagging approaches, say 1/2n-out-of-n without replacement (subagging) and n-out-of-n with replacement (bagging) have been applied in the problem of estimation of the parameters in a multivariate mixtre model. It has been observed by Monte Carlo simulationa that both bagging methods have imperoved the standard deviation of the maximum likelihood estimator of the mixing proportion, whilst the absolute bias increased slightly. In estimating the component distributions, bagging could increase the root mean integrated sguared error when estimating the most probable component.

Keywords:

Multivariate mixture model , maximum likelihood estimation , EM algorithm , Bagging , Subagging