0
$\begingroup$

I am doing load forecasting using SVR(kernel='rbf').How can I understand which is the best value for parameters C, epsilon and gamma?Thanks.

$\endgroup$
4
$\begingroup$

It looks like that you are using scikit-learn. In this case use Grid Search Cross Validation or Randomized Search Cross Validation to find the best parameters.

sklearn.grid_search.GridSearchCV(estimator, param_grid, scoring=None, cv=None, ...)

In these approaches you basically loop over possible sets of your parameters, specified via param_grid. For each set you perform a cross-validation (default is 3-fold, but you can specify it via cv parameter). Cross-validation gives you the mean value and deviation of your 'scoring' parameter (e.g. 'accuracy' for classification, 'r2' for regression). The set of parameters with the best 'scoring' is the winner.

See example here. It also show how to output not only the best set of parameters, but also "top-N". I find it very useful in building intuition about my model.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.