Pelican Optimization Algorithm-based ANFIS for Bolstered Electricity Usage Prediction

Reliable electricity supply is essential for global prosperity, necessitating accurate electricity load forecasts from utilities and policymakers. Conventional prediction methods often fall short, driving a surge in the application of machine learning (ML)-based modeling tools. This paper aims to develop a hybrid model combining the Pelican Optimization Algorithm (POA) and Adaptive Neuro-Fuzzy Inference System (ANFIS) for predicting electricity consumption in a southwestern Nigerian region. Meteorological data from the study area served as inputs, while electricity consumption was the output variable. Evaluated using five performance metrics, the POA-based ANFIS exhibited superior performance, achieving Root Mean Square Error (RMSE) of 1314.7, Mean Absolute Percentage Error (MAPE) of 11.1460, Mean Absolute Relative Error (MARE) of 0.1115, and Coefficient of Variation of Root Mean Square Error (CVRMSE) of 13.0144. The research showcases the promising capabilities of the suggested model as a dependable instrument for predicting energy consumption.