DEV Community

Sajjad Rahman
Sajjad Rahman

Posted on

✅ MCQs

Q1. Linear Models

Which statement about linear models is correct?

A. They always produce nonlinear decision boundaries
B. They use a hyperplane as a decision surface
C. They maximise classification accuracy directly
D. They ignore input features

Answer: B
👉 From PDF: “The decision surface is a hyperplane.”


Q2. Sum of Squared Errors (SSE)

What is the goal when training a linear model?

A. Maximise likelihood
B. Minimise SSE
C. Maximise distance between points
D. Minimise number of features

Answer: B
👉 Linear models minimise SSE


Q3. Effect of Outliers

What is true about samples far from the hyperplane?

A. They have no effect
B. They are ignored
C. They have a stronger effect on the model
D. They reduce variance

Answer: C
👉 “Samples far away from hyperplane have a stronger effect.”


Q4. Linear Separability

When are data points typically linearly separable?

A. When ( n \gg d )
B. When ( n \le d + 1 )
C. Always
D. Never

Answer: B
👉 From PDF: separable when ( n \le d + 1 )


Q5. KNN Prediction

What does KNN use to classify a new data point?

A. Mean of all points
B. Nearest neighbours
C. Gradient descent
D. Covariance matrix

Answer: B
👉 Based on nearest neighbours


Q6. Role of K in KNN

What is the effect of choosing a small K?

A. Smooth decision boundary
B. Linear boundary
C. Highly flexible / irregular boundary
D. No effect

Answer: C
👉 Small K → complex, irregular decision surfaces


Q7. Distance Metrics

Which is NOT mentioned as a valid distance in KNN?

A. Euclidean distance
B. Manhattan distance
C. Hamming distance
D. Fourier distance

Answer: D
👉 Others are listed in PDF


Q8. Curse of Dimensionality

What happens as dimensionality increases?

A. Less data is needed
B. Nearest neighbours become more meaningful
C. Required training data increases exponentially
D. Models become simpler

Answer: C
👉 “Required amount of training examples increases exponentially”


Q9. Cross-Validation

What is the purpose of cross-validation?

A. Increase training error
B. Evaluate model performance on unseen data
C. Remove noise
D. Reduce dimensionality

Answer: B
👉 Used for validation/testing


Q10. Accuracy Definition

Accuracy is defined as:

A. TP / FP
B. (TP + TN) / Total
C. FN / Total
D. TP / (TP + FN)

Answer: B
👉 Standard accuracy formula

Top comments (0)