A novel combined model for prediction of daily precipitation data using instantaneous frequency feature and bidirectional long short time memory networks

In the developing world, to learn nature better, to get the maximum benefit from nature is being studied. Meteorological events constantly affect human life. The occurrence of excessive precipitation in a short time causes important events such as floods. However, in case of insufficient precipitation for a long time, drought occurs. In recent years, significant changes in precipitation regimes have been observed and these changes cause socioeconomic and ecological problems. Therefore, it is of great importance to correctly predict and analyze these variables. In this study, reliable and accurate precipitation forecasting model is proposed. Ensemble of instantaneous frequency (IF) Bidirectional Long Short Time Memory Networks (biLSTM) model was employed for the aim of forecasting of daily precipitation data. To compare the performance of biLSTM model, Long Short Time Memory Networks (LSTM) and Gated Recurrent Unit (GRU) model was applied for forecasting of daily precipitation data. The performance of the proposed IF-biLSTM model was evaluated using Mean absolute error (MAE), Mean square Error (MSE), Correlation Coefficient (R) and Determination Coefficient (R2) performance parameter. According to numerical results, IF-biLSTM model has the best forecasting performance for daily precipitation data. Especially six ahead precipitation forecasting is noteworthy.