Hybrid of ARIMA and SVMs for Short-Term Load Forecasting

摘要

Short-term load is a variable affected by many factors. It is difficult to forecast accurately with a single model. Taking advantage of the autoregressive integrated moving average (ARIMA) to forecast the linear basic part of load and of the support vector machines (SVMs) to forecast the non-linear sensitive part of load, a method based on hybrid model of ARIMA and SVMs is presented in this paper. It firstly uses ARIMA to forecast the daily load, and then uses SVMs, which is known for the great power to learn and generalize, to correct the deviation of former forecasting. Applying this hybrid model to a large sample prediction, the results show that it achieves the forecasting accuracy and has very good prospective in applications. So it can be used as a new load forecasting method.