Analysis of electricity consumption is important because it concerns the resilience of a country. Data mining techniques are needed that can handle fluctuations in these data, one of which uses SVR. The best parameter search for SVR is the most important for forming the model. In this paper, we will discuss in detail the combination of SARIMA-SVR and Firefly algorithms. The firefly algorithm is one of the metaheuristic techniques that provides accuracy which proven by the value of MAPE and RMSE.
Keywords electrical load, SVR, Firefly, Metaheuristics, Forecasting