Advancing Freshwater Lake Level Forecast Using King’s Castle Optimization with Training Sample Adaption and Adaptive Neuro-Fuzzy Inference System

Abstract

This study presents a novel method for more accurate forecasting freshwater Lake Levels with complex fluctuation patterns due to multiple anthropogenic demands and climate factors. The new method employs the mighty King’s Castle Optimization (KCO) with Training Sample Adaption (TSA) and Adaptive Neuro-Fuzzy Inference System (ANFIS) to develop a novel hybrid KCO-TSA-ANFIS model. The performance of the new KCO-TSA-ANFIS Lake water level forecast model is tested on the monthly water levels of Lake Van, in Turkey, showing significantly improved accuracy in model forecasts compared with the regular ANFIS model. By comparing the Root Mean Square Error (RMSE) results, it can be concluded that the KCO-TSA-ANFIS method has 71% higher performance than the simple ANFIS method.