Can anyone assist on how SVR is used in predicting values in a time series data considering the fact that training and testing in SVR is done sometimes randomly. I want to know the difference in how SVR is applied to time series data and cross-sectional data.