We test the performance of different hyper-parameters using cross-validation.
Let's say, the hyper-parameter H1 had the highest average score in cross-validation.
Now, we will take H1 and then train a new model but this time we will use complete training data. No test and train testing is required this time as we have already proven the stability/performance of hyper-parameter H1 using cross validation.
This is how you will get to the final model, right?
Hyperparameter tuning and cross-validation:
We test the performance of different hyper-parameters using cross-validation.
Let's say, the hyper-parameter H1 had the highest average score in cross-validation.
Now, we will take H1 and then train a new model but this time we will use complete training data. No test and train testing is required this time as we have already proven the stability/performance of hyper-parameter H1 using cross validation.
This is how you will get to the final model, right?