DataX Quiz: Machine Learning Mastery for Data Professionals

1. Gradient descent updates model weights by moving in which direction?

Direction of steepest ascent

Direction of steepest descent

Random direction

Perpendicular to loss surface

Submit

5. One-hot encoding is used to transform which type of variable?

Continuous variables

Categorical variables

Time series data

Numerical features

6. Which sampling technique ensures equal representation of classes?

Random sampling

Stratified sampling

Systematic sampling

Cluster sampling

7. Precision measures the proportion of what?

Correct predictions among all positive predictions

Correct predictions among actual positives

Correct predictions overall

Negative predictions correctly identified

8. Which distance metric is most suitable for K-Means clustering?

Cosine similarity

Euclidean distance

Manhattan distance

Hamming distance

9. What is the purpose of a validation set in model development?

Train the final model

Tune hyperparameters

Report final performance metrics

Detect data leakage

10. Which technique handles missing data by replacing with column mean?

One-hot encoding

Mean imputation

Forward fill

Drop rows

11. Which algorithm is best suited for binary classification problems?

Logistic Regression

K-Means Clustering

Principal Component Analysis

Linear Regression

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

Linear

ReLU

Sigmoid

Softmax

13. What is the main advantage of using Principal Component Analysis (PCA)?

Improves model accuracy

Reduces dimensionality while preserving variance

Handles missing values

Increases computational speed

14. The confusion matrix is primarily used to evaluate which type of problem?

Regression

Clustering

Classification

Dimensionality reduction

15. Which ensemble method combines weak learners sequentially?

Bagging

Boosting

Stacking

Voting

16. What does regularization help prevent in machine learning models?

Underfitting

High variance and overfitting

Class imbalance

Missing values

17. Feature scaling is essential for which algorithm?

Decision Trees

K-Nearest Neighbors

Gradient Boosting

Random Forest

18. What is the primary purpose of cross-validation?

Reduce computational cost

Estimate model generalization performance

Eliminate outliers

Increase training speed

19. Which metric is most appropriate for imbalanced datasets?

Accuracy

F1-Score

Precision alone

Recall alone

20. What does overfitting occur when?

Model complexity exceeds training data size

Training loss decreases below validation loss

Model learns noise instead of patterns

All of the above

DataX Quiz: Can You Master Machine Learning for Data Pros?