Electric Load Forecasting in Maros City Based on Extreme Learning Machine (ELM)

Abstract

Electrical load forecasting is one way to reduce the risk of unstable electricity supply. In writing this journal, electrical load forecasting is carried out using an artificial neural network using the Extreme Learning Machine (ELM) method. Extreme Learning Machine (ELM) is a new learning method in artificial neural networks with a single layer feedforward neural network model. The accuracy of the Extreme Learning Machine (ELM) method can be calculated using the Mean Absolute Percentage Error (MAPE). Based on the implementation carried out From the electricity load data for the city of Maros, it is known that of the two extreme learning machine activation functions that were simulated (linear and logsig), the linear activation function provides better daily electrical load forecasting results with a MAPE value of 5.44%.