Font size
WorksheetsRegression Classification
Total questions: 38
Worksheet time: 19mins
When to use the accuracy metric?
When your dataset is balanced
When your dataset is unbalanced
When your dataset is huge in size
None of the above
What are the evaluation metrics used for classification?
Precision
recall
f1-score
All of the above
Which of the following are classification algorithms?
Logistic Regression
Linear Regression
KNN
Support Vector Machine
In Multi-Class Classification
There are 2 unique labels
There are more than 2 unique labels
There is only one label
None of the above
Logistic regression is
Its regression algorithm
Its binary classification algorithm
Its multiclass classification algorithm
None of the above
When to use accuracy as an evaluation metric?
When the dataset is imbalanced
When the dataset is balanced
When precision is more important than recall
None of the above
Which of the following metrics is most suitable for imbalanced datasets?
Accuracy
Precision
Recall
F1-Score
Precision measures:
Correctly predicted positives / Total predicted positives
Correctly predicted positives / Total actual positives
True positives + False negatives
Total correct predictions / Total predictions
Recall measures:
Correctly predicted positives / Total actual positives
Correctly predicted negatives / Total actual negatives
False positives / True negatives
Precision × Recall
F1-score is the:
Harmonic mean of precision and recall
Arithmetic mean of precision and recall
Weighted average of precision and recall
None of the above
In binary classification, the ROC curve plots:
Precision vs Recall
True Positive Rate vs False Positive Rate
Accuracy vs Error rate
Sensitivity vs Specificity
Area under ROC curve (AUC) closer to 1 means:
Poor classifier
Random classifier
Strong classifier
None of the above
Confusion matrix diagonal values represent:
Incorrect classifications
Correct classifications
Misclassifications
False predictions only
Logistic Regression is:
Linear regression for classification
Regression for continuous values
Always multi-class
None of the above
Logistic Regression output lies between:
-∞ to +∞
0 to 1
-1 to +1
None of the above
In Logistic Regression, sigmoid function is used to:
Scale features
Map input to probability between 0 and 1
Improve accuracy
Regularize model
Logistic regression can be extended to multi-class classification using:
One-vs-One (OvO)
One-vs-Rest (OvR)
Softmax regression
All of the above
Which algorithm uses decision boundary based on maximizing margin?
Logistic Regression
SVM
KNN
Naïve Bayes
Kernel trick in SVM is used for:
Reducing dataset size
Handling missing data
Transforming non-linear data to higher dimensions
Reducing overfitting
In KNN, larger value of K leads to:
High variance
High bias
High overfitting
More complex decision boundary
Which metric is most important for medical diagnosis (cancer detection)?
Accuracy
Precision
Recall
Specificity
Precision is also known as:
Sensitivity
Specificity
Positive Predictive Value
Negative Predictive Value
Recall is also known as:
Sensitivity
Specificity
Accuracy
False Positive Rate
Which of the following can handle non-linear decision boundaries effectively?
Logistic Regression
Linear SVM
KNN
Naïve Bayes
Which classification algorithm is based on Bayes’ theorem?
KNN
SVM
Naive Bayes
Logistic Regression
In Naive Bayes, “naive” assumption is:
All features are dependent
All features are independent
Features are correlated
Labels are independent
Which of the following is true about F1-score?
High F1-score means balance between precision & recall
Low F1-score means poor recall
F1-score is useful in imbalanced data
All of the above
Which classifier works by voting mechanism?
Decision Trees
KNN
Logistic Regression
SVM
Which algorithm is most interpretable for classification?
Logistic Regression
SVM with RBF kernel
KNN
Neural Networks
Overfitting in KNN occurs when:
K is very small
K is very large
Features are normalized
Dataset is large
Which metric penalizes false negatives more?
Precision
Recall
Accuracy
Specificity
Which classification algorithm is best suited for text classification?
Logistic Regression
Naive Bayes
SVM
KNN
Which classification metric is threshold-independent?
ROC-AUC
Accuracy
Precision
Recall
Which classification algorithm is most prone to overfitting without regularization?
Logistic Regression
Decision Tree
KNN
Naive Bayes
The decision boundary of logistic regression is always:
Linear
Non-linear
Circular
Random
Softmax function in multi-class classification ensures:
Each class gets probability between 0 and 1
Sum of probabilities = 1
One-vs-rest classification
Both a and b
Which of the following is most sensitive to outliers?
Logistic Regression
KNN
Naïve Bayes
SVM
Which metric is used to compare the Precision and Recall trade-off?
ROC curve
Precision-Recall curve
Confusion Matrix
Specificity
