Optimum estimation of missing values in randomized complete block design by genetic algorithm

Missing data are a part of almost all research, and we all have to decide how to deal with it from time to time. There are a number of alternative ways of dealing with missing data. The problem of handling missing data has been treated adequately in various real world data sets. Several statistical methods have been developed since the early 1970s, when the manipulation of complicated numerical calculations became feasible with the advancement of computers. The purpose of this research is to estimate missing values by using genetic algorithm (GA) approach in a randomized complete block design (RCBD) table and to compare the computational results with three other methods, namely, particle swarm optimization (PSO), Artificial Neural Network (ANN), approximate analysis and exact regression method. Furthermore, 30 independent experiments were conducted to estimate missing values in 30 RCBD tables by GA, PSO, ANN, exact regression and approximate analysis methods. Computational results indicated that the best answer (in the last 10-chromosome population) obtained by GA is frequently the same as the missing value, with the mean value being close to the missing observation. It is concluded that GA provides much better estimation than the other methods. The superiority of GA is shown through lower error estimations and also Pearson correlation experiment. Therefore, it is suggested to utilize GA approach of this study for estimating missing values for RCBD.