Application of Input Variables Selecting Method for Support Vector Machine Model

It is very important to select input variables when the support vector machine forecasting model is proposed. The input variables selection for short-term load forecasting is relevant to the performance of support vector machine forecasting. By using the correlation coefficient idea on input variables selection for support vector machine short-term load forecasting, a systemic and operable method for input variables sets selection is first proposed. An example of short-term load forecasting is given. The result shows that a more preferable input variables set can be obtained, and the forecasting errors are smaller, which validates that the method is effective