Abstract: With the rapid growth of computational domains, like bioinformatics finance, engineering, biometrics and neuro-imaging, emphasizes the necessity for analysing high dimensional data. Many real world data sets may contain hundreds or thousands of features. A common problem that may occur in many knowledge classification systems is presence of incomplete data samples with missing or unknown attribute values which will downside the quality of classification results. Due to the presence of a large segment of missing values in the datasets, refined multiple imputation methods are required to estimate the missing values, so that, a fair and more consistent analysis can be achieved. This study is implemented in Horton works Sandbox on Microsoft Azure. Three imputation (MI) methods are employed, i.e., imputation by mean, imputation by predictive mean and imputation by additive LASSO. Results show that imputations by additive LASSO is the preferred Multiple Imputation (MI) method.