Trees

Calculate the Gini index for a sample composed by 4 red and 0 blue.

• 0

• 1

• 4

• 50

• 100

Calculate the Gini index for a sample composed by 2 red and 2 blue.

• 0.2

• 0.5

• 1

• 2

• 50

You are growing a Decision tree for regression on a dataset with 10,000 points. How many leaves do you expect to have?

• 0

• 5000

• 10000

• 20000

Your model is overfitting. How do you fix that?

• I increase all the hyperparameters

• I decrease all the hyperparameters

• I increase the min_* hyperparameters and decrease the max_* hyperparameters

• I decrease the min_* hyperparameters and increase the max_* hyperparameters

If you train the same Decision Tree Classifier 1000 times on a data set. Are these classifiers perfectly independent to each other?

• Yes

• No

• Sometimes

• I don't know

Setting n_estimators to a very large number would produce

• Overfitting

• Stable predictions

• Underfitting

• I don't know

Which is the algorithm that can be trained faster?

• Random Forest

• Extra-Tree

• I don't know

Which is the algorithms that is less prone to overfitting?

• Random Forest

• Extra-Tree

• I don't know

Setting n_estimators to a very large number would produce

• Overfitting

• Stable predictions

• Underfitting

• I don't know

A feature (gender: male/female) is very predictive of the response. However, once it has been used, it might not be used again because there may be other features (age: ranging from 0 to 100) which are more predictive at later stages of modelling.

• high gain, low cover, low weight

• low gain, high cover, low weight

• high gain and cover, but low weight

• I don't knoe