Application of Sequential Regression Multivariate Imputation Method on Multivariate Normal Missing Data

Missing values means the absence of data items for an observation that can result in the loss of certain information. During surveys, there are often missing values or missing data because there are likely respondents who cannot answer the question or do not want to answer the question. One way to handle missing values can be done by imputation, which is the process of filling or replacing missing values in the dataset with possible values based on information obtained in the dataset. This paper will apply the sequential regression multivariate imputation (SRMI) method for imputation of missing values in normal multivariate data. SRMI is a multiple imputation method whose imputation values are obtained from the sequence of regression model, where each variable containing missing values is regressed against all other variables that do not contain missing values as predictor variables. The way to get the value of imputation is to use an iteration approach to draw values from the predictive posterior distribution of the missing values under each successive regression model. the results of the evaluation of imputation quality on simulation data using Root Mean Square Error (RMSE).