DataX Model Evaluation Metrics Overview

1. In clustering evaluation, what does the silhouette score measure?

How similar points within a cluster are compared to points in other clusters

The total number of clusters formed

The computational time required for clustering

The variance within each feature dimension

2. Which approach best prevents data leakage during model evaluation?

Performing all preprocessing and feature engineering before splitting data

Using the same dataset for training and testing

Applying feature scaling before train-test split

Including target variable statistics in feature engineering

3. What is the relationship between sensitivity and specificity?

They represent trade-offs: increasing one often decreases the other

They are always equal in a good model

Specificity depends only on negative class distribution

They measure the same thing using different scales

4. In model evaluation, what is a 'baseline model' used for?

To provide a reference point for comparing model performance

To train the final production model

To eliminate outliers from the dataset

To determine the optimal number of clusters

5. The sensitivity (recall) of a model is 0.95. What does this mean?

95% of actual positive cases are correctly identified

95% of predictions are positive

95% of negative cases are correctly identified

95% of all predictions are accurate

6. What is the primary advantage of using a validation set separate from the test set?

To tune hyperparameters without biasing final test performance estimates

To increase the total amount of training data

To reduce model training time

To eliminate the need for cross-validation

7. In regression evaluation, what does RMSE (Root Mean Squared Error) penalize?

Large errors more heavily than small errors

All errors equally regardless of magnitude

Only positive prediction errors

The number of features in the model

8. What does the term 'generalization error' refer to?

The difference between training error and test error performance

The total number of errors in the training set

The standard deviation of predictions

The error rate of the simplest baseline model

9. Which metric is most suitable for highly imbalanced datasets?

Accuracy alone

F1-score or balanced accuracy

Mean squared error

Pearson correlation coefficient

10. What is the purpose of stratified sampling in train-test splits?

To maintain the same class distribution in training and test sets

To randomly shuffle the data

To remove duplicate records

To normalize feature scales

11. What does the Area Under the Curve (AUC) metric primarily measure in classification models?

The trade-off between true positive rate and false positive rate across thresholds

The total number of correct predictions

The average time required for model training

The number of features used in the model

12. What does the confusion matrix in classification display?

True positives, true negatives, false positives, and false negatives

Only the correct predictions

The feature correlation values

The model's training loss over iterations

13. Which validation approach is best for small datasets?

Leave-One-Out Cross-Validation (LOOCV)

Simple train-test split with 80-20 ratio

Random sampling without replacement

Using the entire dataset for both training and testing

14. What is overfitting in the context of model evaluation?

When a model performs well on training data but poorly on unseen data

When a model uses too many features

When training takes too long to complete

When predictions exceed the range of target values

15. In regression, what does R² (coefficient of determination) represent?

The proportion of variance explained by the model

The average prediction error magnitude

The correlation coefficient between features

The residual sum of squares only

16. What is the F1-score used for?

Measuring the harmonic mean of precision and recall

Computing the variance of model predictions

Calculating feature importance scores

Determining optimal learning rate for neural networks

17. A model has high recall but low precision. What does this indicate?

The model correctly identifies most positive cases but has many false positives

The model is very accurate overall

The model fails to capture any positive instances

The model has perfect balance between sensitivity and specificity

18. What does precision measure in binary classification?

The proportion of positive predictions that are actually correct

The proportion of actual positives correctly identified

The total accuracy across all classes

The ratio of false positives to true negatives

19. Which metric is most appropriate for evaluating a regression model with outliers?

Mean Squared Error (MSE)

Median Absolute Error (MAE)

R-squared (R²)

Mean Absolute Percentage Error (MAPE)

20. In cross-validation, what is the primary purpose of dividing data into k folds?

To reduce the size of the training dataset

To estimate model performance using multiple train-test splits

To ensure all features are equally weighted

To eliminate the need for a separate test set

DataX Model Evaluation Metrics Overview Quiz