- We have learned what is Stacking, so now try to build Stacking Classifier.
- For this you need to recap your numpy skills, we have loaded the necessary data and packages for you.
Note :- Before writing your solution define the parameters and model as mentioned below and the start your function to implement on these model.
- You will take three models, one BaggingClassifier with logistic regression and other BaggingClassifier's with two decision tree and build stackingclassifier using stacking_clf function with meta_classifier as logistic regression.
- You will use random state 9 for each model and additionally add parameter max_depth = 9 for third model (decision tree).
- You will use n_estimators, max_samples to 100 and bootstrap=True, oob_score set to True in each of the BaggingClassifier.
- Runs a loop and fits each model onto training set which predicts on the training dataset.
- Their will be two stage, First stage will train on the training set and will convert train set to (429,6) numpy array,do the same with the test set and convert to (185,6) numpy array.
- Second stage will be fitting with these newly created numpy array with the meta classifier and perdict the output.
| Parameter | dtype | argument type | default value | description |
|---|---|---|---|---|
| X_train | DataFrame | compulsory | Dataframe containing feature variables for training | |
| X_test | DataFrame | compulsory | Dataframe containing feature variables for testing | |
| y_train | Series/DataFrame | compulsory | Training dataset target Variable | |
| y_test | Series/DataFrame | compulsory | Testing dataset target Variable | |
| model | compulsory | Contains three model that have mention |
| Return | dtype | description |
|---|---|---|
| Accuracy of the model | float | Accuracy of the model for test dataset |
Hint :
- You can use accuracy_score to check the scores
- Function to use np.concatenate, for loop to combine the results of several models into one data set.