ตัวแบบการถดถอยร่วมกับการแปลงแบบเวฟเล็ตและตัวแบบโครงข่ายประสาทเทียมร่วมกับการแปลงแบบเวฟเล็ตสำหรับพยากรณ์ปริมาณน้ำท่าในลุ่มน้ำทะเลสาบสงขลา

Abstract

This study aimed to compare the forecasting efficiency of daily runoff in Songkhla Lake Basin by Wavelet Regression Model and Wavelet Neural Network Model. 70% of daily rainfall and runoff were randomly selected for calibration of models. Data from December 26", 2018 to October 31, 2019 were used for validation of models. The performance of the models was evaluated based on these statistics; Coefficient of Determination (R), Nash-Sutliffe Efficiency Coefficient (ENs) and Root Mean Square Error (RMSE). The results showed that Wavelet Regression Model and Wavelet Neural Network Model were able to explain the variation in Songkhla Lake Basin's runoff equally at 99.99%. The value of Nash-Sutcliffe Efficiency Coefficient (Ens) and Root Mean Square Error (RMSE) of Wavelet Neural Network Model was better. Hence Wavelet Neural Network Model was performed more efficient than Wavelet Regression Model.