Practical ML Quiz 1: Forests and Features

A colorful illustration of decision trees, random forests, and gradient boosting machines in a vibrant, educational style, showcasing various machine learning concepts and their applications.

Practical ML Quiz: Forests and Features

Test your knowledge on machine learning concepts, focusing on forests and features with our comprehensive quiz. Whether you're just beginning or looking to refine your expertise, this quiz covers essential topics related to Random Forests, GBDT, and model evaluation.

11 engaging questions to challenge your understanding
Multiple choice and checkbox formats for diverse learning
Designed for machine learning enthusiasts and practitioners

11 Questions3 MinutesCreated by LearningTree234

We've trained a RandomForest model with N trees. Consider two cases:

1. Drop the first tree

2. Drop the last tree

Then compare models performance on the train set. Select the right answer:

In the case1 performance will drop more than in the case2

In the case1 performance will be roughly the same as in the case2

In the case1 performance will drop less than in the case2

We've trained a GBDT (gradient boost decision tree) with N trees and a fairly high learning rate. Consider two cases:

1. Drop the first tree

2. Drop the last tree

We then compare models performance on the train set. Select the right answer.

In the case1 performance will drop more than in the case2

In the case1 performance will be roughly the same as in the case2

In the case1 performance will drop less than in the case2

Consider the two cases:

1. Fit two RandomForestClassifiers 500 trees each and average their predicted probabilities on the test set.

2. Fit one RandomForestClassifier with 1000 trees and use it to get test set probabilities.

All other hyperparameters are the same for all models. Select the right answer.

The quality of predictions in the case1 will be roughly the same

The quality of predictions in the case1 will be higher

The quality of predictions in the case1 will be lower

From left to right, what model was most probably used to produce such decision surface?

Decision Tree, Random Forest, K Nearest Neighbours, Linear Regression

Decision Tree, K Nearest Neighbours, Random Forest, Linear Regression

K Nearest Neighbours, Decision Tree, Random Forest, Linear Regression

Linear Regression, K Nearest Neighbours, Random Forest, Decision Tree

Random Forest, Decision Tree, K Nearest Neighbours, Linear Regression

K Nearest Neighbours, Linear Regression, Decision Tree, Random Forest

Select the correct statement about the RandomForest and GBDT models.

Trees in RandomForest can be constructed in parallel

In GBDT trees are independent from each other

Trees in GBDT can be constructed in parallel

In RandomForest each new tree is built to improve the previous trees

Suppose we have a feature with all the values between 0 and 1 except few outliers larger than 1

What can help to decrease outliers' influence on non-tree models? Select ALL that applies

Suppose we fit a tree-based model. In which cases label encoding can be better to use than one-hot encoding?

Select ALL that applies

When categorical feature is ordinal

When the number of categorical features in the dataset is huge

When categorical features have high correlation with the ground truth

Suppose we have a categorical feature and a linear model. We need to somehow encode this feature. Which of the following statements are true?

Depending on the dataset either of label encoder or one-hot encoder could be better

Label encoding is always better than one-hot encoding

One-hot encoding is always better than label encoding

Practical ML Quiz 1: Forests and Features

Practical ML Quiz: Forests and Features

More Surveys