Forecasting Electricity Consumption in Malaysia by Hybrid ARIMA-ANN

Forecasting electricity consumption is of national interest to any country. Electricity forecast is not only required for short-term and long-term power planning activities but also in the structure of the national economy. Electricity consumption time series data consists of linear and non-linear patterns. Thus, the patterns make the forecasting difficult to be done. Neither autoregressive integrated moving average (ARIMA) nor artificial neural networks (ANN) can be adequate in modeling and forecasting electricity consumption. The ARIMA cannot deal with non-linear relationships while a neural network alone is unable to handle both linear and non-linear pattern equally well. This research is an attempt to develop ARIMA-ANN hybrid model by considering the strength of ARIMA and ANN in linear and non-linear modeling. The Malaysian electricity consumption data is taken to validate the performance of the proposed hybrid model. The results will show that the proposed hybrid model will improve electricity consumption forecasting accuracy by compare with other models. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.