I don't understand how parameter sweep is done in the Azure Machine Learning?

Question

In azure ml, if we select a training algorithm(for ex "Two-Class Logistic Regression") we can then have a set of parameters to do a parameter sweep while training. But how can I know how they change the values of parameters in the training?

Answer 1

1. Go to Experiments and search for Parameter Optimization. Open the sample experiment in Studio.

2. In Two-class support Vector Machine, set Create Trainer Mode to Parameter range. You can specify the range according to you. 

3. The left Evaluate model will give you a table of metrics for the combination of parameters that were swept over and the other one will give you the best model with your metrics.