8.4. GridSearchCV and RandomizedSearchCV - Python implementation | Hyperparameter Tuning

. This video is about the python implementation of GridSearchCV and RandomizedSearchCV. These are the two important Hyperparameter tuning in Machine Learning.

Hello everyone! I am setting up a donation campaign for my YouTube Channel. If you like my videos and wish to support me financially, you can donate through the following means:

(No donation is small. Every penny counts)
Thanks in advance!

I am making a "Hands-on Machine Learning Course with Python" in YouTube. I'll be posting 3 videos per week: Monday Evening; Wednesday Evening; Friday Evening.

[juliazhu]

The course is very in-depth. It helps me a lot.

[fun_stuff_and_games]

Hi Siddha, you offer great way to explain things, thanks! Just a question on the steps to follow.
Are the steps the following:

1. Run train_test split and display scores for a number of different models (e.g. random forest, decision tree, svc…) >>> this is from video 8.2

2. Validate the score performance seen in step 1 via cross-validation and pick the best performing model (thinking: It could be that a model that is best in train_test_split, is not the best model when running cross-validation. >>> this is from video 8.2

3. Take the best model (let’s say it’s Random Forest) and perform Grid Search via hyperparameter tuning as you explained in this video. In this case, I don’t have to run GridSearch on all models, as I already defined in steps 1 and 2 which one is the best and I am going to use.

4. Once the best model has been validated, run the model on the entire dataset, meaning on X (instead of X_train, X_test) as we don’t need to test anything anymore (we know which ones are the best ones).

Are these steps correct or am I confusing something?

[prathapcme]

why did you take combinations like 1, 5, 10, 20 ...how to take the values

[armin.falahatkar]

hi
thank you so much
it was helpful
i have a question ? when we use gridsearchCV and use cv=5 on it, it's mean that cross_val_score running and data separate to cv=numbers of folds ? like when using cross_val_score alone ?

[siddharthmishra]

How long you will take to completely upload all the remaining video for this ML course???

[fun_stuff_and_games]

So, what I am basically asking is. When do we use Grid Search? Before cross validation or after cross validation?

[satire]

for regression will the parameters for tunig remain same or change ?

[utkarshtripathi]

Sir could u pls make videos on Real industry data science projects