IS353IA
USN
RV COLLEGE OF ENGINEERING
Autonomous Institution affiliated to VTU
V Semester B.E.
Model Question Paper
DEPARTMENT OF INFORMATION SCIENCE AND ENGINEERING
ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING
(2022 SCHEME) Model Question Paper
Time: 03 Hours Maximum Marks: 100
Instructions to candidates:
1. Answer all questions from Part A. Part A questions should be answered in
first three pages of the answer book only.
2. Answer FIVE full questions from Part B. In Part B question number 2 is
compulsory. Answer any one full question from 3 and 4, 5 and 6, 7 and 8,
and 9 and 10.
PART-A
1 1.1 What is the primary advantage of depth-first search (DFS) over breadth-
first search? 01
1.2 Differentiate between a rational agent and a non-rational agent. 02
1.3 A* search considered optimal when the heuristic is admissible. Give reason 01
1.4 In a two-player game, how does the minimax algorithm ensure that both 02
players play optimally?
1.5 How does simulated annealing differ from hill-climbing when reaching a 01
local optimum?
1.6 In K-means, the objective is to minimize the within each cluster, 01
which is the sum of squared distances between data points and their
cluster centroids.
1.7 In Logistic Regression, the parameters are learned by minimizing the 01
function, typically using gradient descent.
1.8 Bagging improves the stability of a classifier by generating multiple 02
datasets through sampling and training a separate model on each
dataset.
1.9 In a decision tree, the Gini index of a split is 0.5, and in another split, it is 02
0.3. Which split is considered better for classification? Give reason
1.10 What are resubstitution errors ? 02
1.11 When pruning a decision tree, why is it important to balance model 01
complexity and accuracy?
1.12 Ensemble methods combine multiple individual models to create a 01
model that typically outperforms any individual model
1.13 In a Naive Bayes classifier, the likelihood of observing a feature value given 01
a class is calculated using probability.
1.14 A major disadvantage of the Nearest Neighbor Classifier is that it can
become computationally expensive as the number of in the
dataset increases. 01
1.15 To improve the accuracy of K-means clustering, it may be helpful to
apply to the dataset before clustering to normalize the data. 01
PART-B
Explain how a problem-solving agent works. Describe the steps it follows to
a solve a problem, using an example of a typical problem like route planning 6
Traverse the given graph using BFS and DFS and compare the results
2
b
10
Compare and contrast A* search with Greedy Best-First Search. Use
a 10
examples to illustrate their differences.
Apply A* algorithm to the below given graph and find the optimal path
3
b
OR
What are local search algorithms? Identify and explain using examples, the
a 10
key features that distinguish them from classical search methods.
4
Critically evaluate the effectiveness of simulated annealing in finding global
b
optima for complex search spaces. 6
Construct a decision tree for the dataset given using ID3 algorithm
a
5
10
