Comparison of Predictive Statistical Learning Accuracy with Computational Intelligence Methods

Forecasting of high-frequency economic time series data is a complex problem, which has benefited from recent advancements and research in machine learning. To forecast time series data, two methodological frameworks of statistical and computational intelligence modelling are considered. The statistical methodological approach is based on the theory of invertible ARIMA (Auto-Regressive Integrated Moving Average) models. The Computational Intelligence (CI) models are based on neural networks (NN) and Support Vector Machines (SVM). The main work of this study is to compare the predictive accuracy level of the statistical methodological approach with NN and SVM on the large data set. We evaluate statistical ML (Maximum Likelihood) learning method, Back- Propagation (BP) and genetic algorithms (GA) for half-hourly 1-step-ahead electricity demand prediction using Australian electricity data. We showed that all ARIMA, NN, SVM models as prediction methods are reasonable and acceptable for use in forecasting systems that routinely predict values of variables in competitive energy markets.