Distributed Imputation for Random Effects Models with Missing Responses

By adding over-relaxation factor to PXEM (Parameter Expanded Expectation Maximization) method, the MOPXEM (Monotonically Overrelaxed Parameter Expanded Expectation Maximization) method is obtained. Compare it with the existing EM (Expectation-Maximization)-like methods. Then, distribute and process five methods and compare them, achieving good performance in convergence speed and result quality.The philosophy of the package is described in Guo G. (2022) .