Wow, that was a lot to take in! Let’s take this opportunity to implement AdaBoost on a real dataset and solve a classification problem.
We will be using a dataset from UCI’s Machine Learning Repository to evaluate the acceptability of a car based on a set of features that encompasses their price and technical characteristics.
Instructions
Create the base estimator for the AdaBoost classifier in the form a decision stump using DecisionTreeClassifier() and store it in a variable named decision_stump. Recall, that a decision stump is a decision tree with only two leaf nodes.
Print the parameters of the decision stump using the .get_params() method.
Create an AdaBoost classification model with the base_estimator parameter set to decision_stump and n_estimators set to 5. Store the model in a variable named ada_classifier.
Print the parameters of the AdaBoost model using the .get_params() method.
Fit ada_classifier using the training features (X_train) and corresponding labels (y_train).
Predict the classes of the testing dataset (X_test) and store them as an array in a variable named y_pred.
Now we will explore some of the most common evaluation metrics for classification on our trained AdaBoost model.
- Calculate the accuracy and store it in a variable named
accuracy. - Calculate the precision and store it in a variable named
precision. - Calculate the recall and store it in a variable named
recall. - Calculate the f1-score and store it in a variable named
f1.
Remove the comments from the code block to print the evaluation metrics you just stored.
Take a look at the confusion matrix by removing the comments in the following code block.