How to get hold of performance figures for validation_set?

User 3031 | 1/15/2016, 5:12:05 PM

I am taking the ML class on coursera and am playing around with linearregression.create() and how to work with a validation set. I notice that if I set validationset = "auto" or validationset = validata (provided I have created a validata sFrame), and set verbose = True, I can see the model generating performance numbers for both the training and validation sets (trainingrmse and validation_rmse). But how can I explicitly make use of the validation performance numbers? For example, if I ask for model['training_rmse'] or model['trainingloss'] I get the rmse and the RSS for the training data, but I can find no way to get hold of the corresponding validationset numbers, such as model['validation_rmse'] or model['validation_loss'], to be used as variables in subsequent steps.


User 2568 | 1/16/2016, 2:18:46 AM

I asked a the same question here