Missing Data Imputation using Gaussian Copula
This package fits a Gaussian copula model to the incomplete data and impute the missing values using the fitted model. The data input can have continuous, ordinal and binary variables, but categorical variable is not supported. The user can either fit a full rank model (preferred when the number of variables is small) or a low rank model (preferred when the number of variables is large). The user can also conduct online model fitting and missing data imputation for time series data. Mini-batch training and parallelization are also supported to accelerate the offline run.