1. Question 1

Logistic regression predicts:

• ❌ Price

• ❌ Age

• ❌ Weight

• ✅ The class a sample belongs to

Explanation:
Logistic regression is a classification model.

---

2. Question 2

Class vector y represents:

• ❌ Features

• ✅ Discrete class labels

• ❌ Continuous values

• ❌ Bias terms

Explanation:
For binary logistic regression, y ∈ {0, 1}.

---

3. Question 3

If a dataset can be separated by a line, it is:

• ❌ Nonlinear

• ❌ Multiclass

• ❌ Unclassifiable

• ✅ Linearly separable

---

4. Question 4

In w · x + b, the term b is:

• ❌ Feature

• ❌ Weight

• ✅ Bias term

• ❌ Sample value

---

5. Question 5

Function used in logistic regression:

• ❌ ReLU

• ❌ Tanh

• ✅ Sigmoid function

• ❌ Linear function

---

6. Question 6

PyTorch package for quickly building logistic regression models:

• ❌ torch.optim

• ❌ torch.autograd

• ❌ torch.nn.functional

• ✅ torch.nn.Sequential

Explanation:
nn.Sequential allows stacking linear + sigmoid quickly.

---

7. Question 7

Function of nn.Sigmoid():

• ❌ Initialize parameters

• ❌ Create linear model

• ✅ Apply sigmoid activation function

• ❌ Linear transformation

---

8. Question 8

θ in a Bernoulli distribution represents:

• ❌ Probability of failure

• ❌ Variance

• ❌ Standard deviation

• ✅ Probability of success

---

9. Question 9

Likelihood of a sequence of Bernoulli events:

• ❌ Add

• ❌ Divide

• ✅ Multiply probabilities

• ❌ Subtract

Explanation:
Independence assumption → likelihood = product of probabilities.

---

10. Question 10

Purpose of cross-entropy loss:

• ❌ Increase learning rate

• ❌ Maximize misclassification

• ❌ Regularize parameters

• ✅ Minimize misclassified samples

Explanation:
Cross-entropy penalizes wrong predictions strongly → improves classification.

---

🧾 Summary Table

---