The purpose of this dissertation is to evaluate convergence acceleration methods on datasets with large numbers of parameters. The motivation is that EM algorithm slow on London Deaths Data. Try convergence acceleration on a genetic dataset which will have a large number of parameters.