Module-3
1 Consider the dataset:
Age Income Buy
Young High No
Young Low Yes
Old High Yes
Old Low Yes
Predict:Age = Young, Income = High → Buy = ? .Compute priors,likelihoods
and probabilities.
2 A model trained on a dataset gives:
Training Accuracy = 60%
Testing Accuracy = 58%
Identify the type of fitting .Explain the reason and suggest methods to
improve performance .
3 Write a Python program to implement Naïve Bayes and explain each step in
the code.
4 Write a Python program using sklearn to classify student performance
using KNN. Dataset:Features: Study Hours, Sleep Hours and Target:
Pass/Fail
5 A model has:Bias² = 2 , Variance = 6 and Noise = 1 .Compute total error
and suggest whether model should be simplified or made more complex
.Justify your answer
6 Discuss how the Curse of Dimensionality, Bias–Variance Tradeoff, and
Overfitting are related.
7 Consider the dataset as shown in Table in Table [Link] point (3,3)
using K = 3.
Table 5a.
X1 X2 Class
1 1 A
2 2 A
4 3 B
5 5 B
8 Explain bias and variance. Describe the bias–variance tradeoff with
diagram.
9 A machine learning model gives the following accuracies:
Training Accuracy = 98%
Testing Accuracy = 72%
Identify whether the model is correct fitting, overfitting, or underfitting
.Justify your answer and suggest one method to improve the model .
1 Consider the dataset as shown in Table 6a. Predict Email with
0 “Offer=Yes”.Compute priors,likelihoods and probabilities.
Table 6a.
Word “Offer” Spam
Yes Yes
Yes Yes
No No
Yes Yes
No No
1 Explain overfitting and underfitting.
1
1 Explain the Curse of Dimensionality.
2
1 Explain KNN algorithm with [Link] points:
3 (1,1), (2,2), (3,3) → Class A
(6,6), (7,7) → Class B
Classify point (4,4) using K=3.
1 Compare overfitting correctfiitting and underfitting in
4 machine learning.
1 Explain bias–variance tradeoff with diagram.
5
1 A hospital screens 10,000 people:
6
1% have the disease
Test detects 95% of diseased cases
10% of healthy people test positive
How many people will test positive? Among them, how many actually have the disease?
What is the probability that a randomly selected positive patient truly has the disease?
1 Describe the steps involved in building a Naïve Bayes
7 model.
1 Write a Python program to classify whether an email is Spam
8 or Not Spam using Naïve Bayes.
1 Explain correct fitting in machine learning.
9
2 A model gives the following errors:
0
Training Error = 2%
Testing Error = 25%
1. Identify whether the model is overfitting or underfitting
2. Justify your answer
3. Suggest two methods to improve the model
2 Differentiate between: KNN Classification versus KNN Regression
1
2 Explain how the Curse of Dimensionality impacts the
2 performance of the k-nearest neighbors algorithm.
2 Consider the dataset:
3 Weather Play
Sunny No
Sunny Yes
Rainy Yes
Rainy Yes
Overcast No
Using Naïve Bayes, predict whether Play = Yes or No when
Weather = Sunny.
2 Discuss overfitting and bias–variance tradeoff.
4