Options
2014
Journal Article
Title
24-hours demand forecasting based on SARIMA and support vector machines
Abstract
In time series analysis the autoregressive integrate moving average (ARIMA) models have been used for decades and due to their success they have been used in a wide variety of scientific applications. In recent years a growing popularity of machine learning algorithms like the artificial neural network (ANN) and support vector machine (SVM) have led to new approaches in time series analysis. Especially the ability to consider nonlinearities gives these methods an advantage. Though it has been shown by [1, 2] that modelling water demand based on auto-regression gives reasonable results for short term prediction it has been presented that a wide variety of factors like the day of week and meteorological conditions can have an impact on the water consumption [3, 4]. The forecasting model presented in this paper combines an autoregressive approach with a regression model respecting additional parameters. The approach is suitable for the forecasting water demand up to 24 hours in advance. For its use in an online context it can be used for short-term one step prediction. Two modelling approaches are presented which are based on seasonal autoregressive integrated moving average (SARIMA) models and support vector regression (SVR). These models are evaluated with respect to their forecasting performance based on data from a residential district in Berlin.