wayground logo

Free Printable Worksheets

NEW

Font size

S
M
L
XL
Worksheets

Classification & Regression - 2

Total questions: 40

Worksheet time: 20mins

Name
Class
Date
1.

Which metric is most important for spam filtering?

a)

Precision

b)

Recall

c)

RMSE

d)

Accuracy

2.

Which metric is most important for disease screening?

a)

Accuracy

b)

Recall

c)

Precision

d)

Specificity

3.

In model evaluation, when both false positives and false negatives are equally costly, the best metric is:

a)

Accuracy

b)

Precision

c)

Recall

d)

F1-score

4.

F1-score is low if:

a)

Both precision and recall are high

b)

Either precision or recall is very low

c)

Accuracy is perfect

d)

Dataset is balanced

5.

In cases where precision = 1 and recall = 0, the F1-score is:

a)

1

b)

0

c)

0.5

d)

Undefined

6.

F1-score is a type of:

a)

Arithmetic mean

b)

Geometric mean

c)

Harmonic mean

d)

Weighted average

7.

Recall is also known as:

a)

Sensitivity

b)

Specificity

c)

Precision

d)

Accuracy

8.

In a medical test, high recall means:

a)

Most healthy people are correctly identified

b)

Most sick people are correctly identified

c)

Few false positives occur

d)

Accuracy is 100%

9.

Recall can be improved at the cost of:

a)

Reducing accuracy

b)

Reducing precision

c)

Reducing F1-score

d)

Increasing false negatives

10.

Precision focuses on reducing:

a)

False negatives

b)

False positives

c)

True negatives

d)

True positives

11.

In a fraud detection system, high precision means:

a)

Most flagged frauds are actually frauds

b)

All frauds are detected

c)

No fraud is missed

d)

System has high accuracy

12.

Precision is also known as:

a)

Sensitivity

b)

Specificity

c)

Positive Predictive Value (PPV)

d)

Negative Predictive Value (NPV)

13.

Accuracy can be misleading when:

a)

Classes are balanced

b)

Classes are imbalanced

c)

Training dataset is large

d)

Model is regularized

14.

In a dataset with 99% negatives and 1% positives, a model predicting all negatives will have accuracy:

a)

100%

b)

99%

c)

50%

d)

1%

15.

Which evaluation metric is better than accuracy for imbalanced datasets?

a)

Precision and Recall

b)

RMSE

c)

R-squared

d)

Variance Inflation Factor

16.

In KNN, smaller values of k usually lead to:

a)

High bias and low variance

b)

Low bias and high variance

c)

Perfect generalization

d)

Always underfitting

17.

Feature scaling (normalization/standardization) is important for KNN because:

a)

Distance calculation depends on scale of features

b)

It prevents overfitting

c)

It improves accuracy always

d)

It reduces training time

18.

Weighted KNN differs from normal KNN by:

a)

Giving equal weight to all neighbors

b)

Giving more weight to closer neighbors

c)

Ignoring distant neighbors completely

d)

Using logistic regression as backend

19.

The main drawback of KNN is:

a)

It requires complex training phase

b)

It is computationally expensive at prediction time

c)

It cannot handle numerical data

d)

It works only for binary classification

20.

The output of logistic regression before applying the sigmoid function is called:

a)

Odds ratio

b)

Logit

c)

Probability

d)

Likelihood

21.

Which of the following is NOT an advantage of logistic regression?

a)

Easy to implement

b)

Works well with linearly separable data

c)

Handles categorical variables

d)

Automatically captures complex nonlinear patterns

22.

Logistic regression can be extended to multi-class classification using:

a)

Linear regression

b)

Softmax regression

c)

Ridge regression

d)

Lasso regression

23.

Logistic regression assumes that:

a)

Relationship between independent and dependent variables is linear in log-odds

b)

Variables are always categorical

c)

The dependent variable is continuous

d)

No correlation exists in data

24.

Which of these is a limitation of classification models?

a)

They cannot predict probabilities

b)

They may be biased by imbalanced datasets

c)

They do not handle categorical features

d)

They are always linear

25.

In classification, the confusion matrix helps in:

a)

Calculating regression coefficients

b)

Summarizing correct & incorrect predictions

c)

Estimating correlation

d)

Removing multicollinearity

26.

Which of the following metrics is scale-independent?

a)

R-squared

b)

MSE

c)

RMSE

d)

MAE

27.

A model with high R² but poor prediction on new data suffers from:

a)

Multicollinearity

b)

Overfitting

c)

Underfitting

d)

Proper fitting

28.

Cross-validation is mainly used to:

a)

Estimate model’s generalization performance

b)

Remove multicollinearity

c)

Reduce correlation

d)

Increase R² artificially

29.

Mean Absolute Error (MAE) is:

a)

Average of squared errors

b)

Average of absolute errors

c)

Root of squared errors

d)

Ratio of errors

30.

Mean Squared Error (MSE) is preferred because:

a)

Easier interpretation

b)

Penalizes larger errors more

c)

Works only with classification

d)

Ignores variance

31.

Adjusted R-squared increases only if:

a)

A new predictor improves the model

b)

A new predictor is always added

c)

A new predictor decreases fit

d)

Sample size reduces

32.

In hypothesis h(x) = β₀ + β₁x, β₁ measures:

a)

Intercept

b)

Average of y

c)

Effect of x on y

d)

Error variance

33.

Residuals in regression represent:

a)

Predicted values

b)

Difference between actual and predicted values

c)

Model coefficients

d)

Independent variables

34.

A fitted regression line is also called:

a)

Hypothesis function

b)

Decision boundary

c)

Error function

d)

Correlation line

35.

One method to handle multicollinearity is:

a)

Adding more correlated variables

b)

Removing highly correlated variables

c)

Ignoring the problem

d)

Increasing p-value threshold

36.

When multicollinearity is high, regression coefficients:

a)

Are very stable

b)

May become unreliable and unstable

c)

Always decrease

d)

Are unaffected

37.

Which of these problems is addressed using regression?

a)

Predicting sales revenue

b)

Classifying spam emails

c)

Grouping customers into clusters

d)

Reducing dimensionality

38.

What is a common issue when using accuracy as a metric for imbalanced datasets?

a)

It requires large datasets

b)

It can be misleading

c)

It is always reliable

d)

It ignores false positives

39.

Which metric is primarily used to evaluate the performance of a model in a binary classification problem?

a)

Mean Squared Error

b)

Adjusted R-squared

c)

F1-score

d)

R-squared

40.

In the context of regression analysis, what does a high R-squared value indicate?

a)

The model is overfitting

b)

The model is linear

c)

The model explains a large proportion of variance

d)

The model has high bias