Data Imputation with Genetic Algorithm and Multiple Linear Regression for Improving Performance of Prediction Model

Prediction model is used to forecast or predict value from dataset. But one of the most common problems in training prediction model is there are missing values in datasets. Problem is usually managed by two methods for solving this problem. First is ignoring, but it reduces the predictive model's performance because of the data that was cut off may be important. Another method is replacing the missing values or data imputation. Benefit of imputation is it still keep all of data. It means an important data will not loss. Therefore, most researchers offer an imputation method for solving this problem. In the past most researches are proposed algorithm that trying to recover the original data, but main object of using prediction model is accuracy of prediction. Algorithm is based on Genetics Algorithm and Multiple Linear Regression is create for improving performance of prediction model.