Question 1

In neural networks, what type of layer is responsible for reducing the dimensionality of the input data?

Options :

A : Hidden

B : Pooling

C : Output

D : Input

Answer: B

Question 2

What is the primary purpose of the CRISP-DM (Cross-Industry Standard Process for Data Mining) framework in the context of data science projects?

Options :

A : To provide a structured approach to the entire data science life cycle

B : To optimize model performance and accuracy

C : To facilitate collaboration among team members

D : To automate data preprocessing and feature engineering

Answer: A

Question 3

In a decision tree, what does a split represent?

Options :

A : A decision based on a feature and a threshold that divides the data into two or more subsets

B : A pruning technique to reduce overfitting

C : The output of the decision tree algorithm

D : A data point in the dataset

Answer: A

Question 4

What principle emphasizes the importance of integrating data science workflows with development and operations processes to streamline deployment and maintenance?

Options :

A : MLOps

B : Data replication

C : DevOps

D : Continuous integration/continuous deployment

Answer: A

Question 5

In machine learning, what technique is commonly used for handling class imbalance?

Options :

A : Oversampling the minority class

B : All of the above

C : Undersampling the majority class

D : SMOTE

Answer: B