Data Science And AI Quiz

1. What is the primary goal of supervised learning?

Find patterns in unlabeled data

Predict output values using labeled training data

Reduce data dimensionality

Cluster similar observations together

2. Which preprocessing step removes or handles missing values in a dataset?

Normalization

Imputation

Feature engineering

Standardization

3. What does overfitting mean in machine learning?

Model performs well on training data but poorly on test data

Model fails to capture patterns in training data

Model uses too few features

Model requires excessive computational resources

4. Which metric is best for evaluating a binary classification model with imbalanced classes?

Accuracy

Precision and recall

Mean squared error

R-squared

5. In k-means clustering, what does k represent?

Number of features

Number of clusters

Number of training samples

Learning rate parameter

6. What is the purpose of cross-validation in machine learning?

Reduce training time

Assess model performance on unseen data

Increase model complexity

Balance feature importance

7. Which activation function is commonly used in hidden layers of neural networks?

Linear

ReLU

Softmax

Sigmoid

8. What does the term 'feature scaling' accomplish?

Removes outliers from data

Transforms features to comparable ranges

Selects the most important features

Increases the number of features

9. In a confusion matrix, what does true positive represent?

Correctly predicted negative cases

Correctly predicted positive cases

Incorrectly predicted positive cases

Incorrectly predicted negative cases

10. Which algorithm is used for dimensionality reduction?

Principal Component Analysis

Random Forest

Naive Bayes

K-Nearest Neighbors

11. What is the purpose of regularization in machine learning?

Increase model complexity

Prevent overfitting by penalizing large weights

Speed up training

Improve feature selection

12. Which type of neural network is best for sequential data like time series?

Convolutional Neural Network

Recurrent Neural Network

Fully connected network

Autoencoders

13. What does gradient descent do in model training?

Selects features for the model

Iteratively reduces loss by updating weights

Splits data into train and test sets

Normalizes input features

14. In ensemble learning, what is the benefit of combining multiple models?

Reduces training time significantly

Improves prediction accuracy and robustness

Eliminates the need for feature scaling

Requires less computational power

15. Which loss function is typically used for multiclass classification?