Online Forecasting of Time Series Using Incremental Wavelet Decomposition and Least Squares Support Vector Machine

Time series is widely concerned in industry engineering, finance, economy, traffic and many other fields. Forecasting of time series is an important work and online forecasting is necessary in some real time application. An efficient method for forecasting of time series using wavelet transform and least squares support vector machine (LS-SVM) is presented, which can provide high accuracy and cost less time. Sliding window model is used to follow the data changing, and incremental algorithms for wavelet decomposition is used to save time. Simulation experiment using real power load dataset show the effectiveness of proposed method.