- Growing the tree beyond a certain level of complexity leads to overfitting
- In our data, age doesn’t have any impact on the target variable.
- Growing the tree beyond Gender is not going to add any value. Need to cut it at Gender
- This process of trimming trees is called Pruning
Pruning to Avoid Overfitting
- Pruning helps us to avoid overfitting
- Generally it is preferred to have a simple model, it avoids overfitting issue
- Any additional split that does not add significant value is not worth while.
- We can use Cp – Complexity parameter in R to control the tree growth
#We will rebuild a new tree by using above data and see how it works by tweeking the parameteres dtree = tree.DecisionTreeClassifier(criterion = "gini", splitter = 'random', max_leaf_nodes = 10, min_samples_leaf = 5, max_depth= 5) dtree.fit(X_train,y_train)
DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=5, max_features=None, max_leaf_nodes=10, min_samples_leaf=5, min_samples_split=2, min_weight_fraction_leaf=0.0, presort=False, random_state=None, splitter='random')
predict3 = dtree.predict(X_train) print(predict3)
[1 1 0 0 0 1 1 1 1 0 0 1 0 0]
predict4 = dtree.predict(X_test) print(predict4)
[1 1 0 0 0 1]
#Accuracy of the model that we created with modified model parameters. score2 = dtree.score(X_test, y_test) score2