Comparison of Artificial Neural Network (ANN) and Other Imputation Methods in Estimating Missing Rainfall Data at Kuantan Station

Abstract

Daily rainfall data could be considered as one of the basic inputs in hydrological (e.g. streamflow, rainfall-runoff, recharge) and environmental (e.g. crop yield, drought risk) models as well as in assessing the water quality. In Malaysia, the number of rain gauge stations with complete records for a long duration is very scarce. The occurrence of missing values in rainfall data is mainly due to malfunctioning of equipment and severe environmental conditions. Thus, the estimation of rainfall is needed, whenever the missing data happened at the principal rainfall station. In this study, daily rainfall data from eight meteorological stations located in Pahang state are considered and Kuantan is selected as the target station. The main purposes of this study is to compare the performance of the imputation methods by using Artificial Neural Network method (ANN), Bootstrapping and Expectation Maximization Algorithm method and Multivariate Imputation by Chained Equations method (MICE). Missing rainfall data has been generated randomly for Kuantan station with 5%, 10% and 15% of missingness. The three methods are compared based on Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Coefficient of Determination (R2). The findings concluded that Artificial Neural Network (ANN) is found to be the best imputation method for this study, followed by Multiple Imputation by Chained Equation (MICE) and Bootstrapping and Expectation Maximization Algorithm method.