A Hybrid Support Vector Regression for Time Series Forecasting

The study applies a novel neural network technique, hybrid support vector regression (SVR), to forecast values of the turbo-generator vibration in time series. The simulation experiment results showed that the hybrid model is superior to the individual models for the test. Most of the individual models evaluated showed poor ability to detect directional change. This problem can be overcome with the use of the hybrid model. Besides superior turning point detectability, the hybrid model could achieve superior predictive performance and showed promising results. Therefore, the results suggested that the proposed hybrid model is typically a reliable forecasting tool for application within the forecasting fields of time series.