End-Term

Question 1

What is Machine Learning?

Answer 1

A field of AI that enables systems to learn from experience without being explicitly programmed.

Question 2

Who defined ML as ‘computers learning from experience E with respect to tasks T and performance measure P’?

Answer 2

Tom Mitchell

Question 3

Name the three types of Machine Learning.

Answer 3

Supervised learning, Unsupervised learning, Reinforcement learning

Question 4

What is supervised learning?

Answer 4

Predicting outcomes based on labeled training data

Question 5

What are examples of supervised learning?

Answer 5

Handwriting recognition, stock prediction, disease diagnosis

Question 6

What is unsupervised learning?

Answer 6

Finding patterns or groupings in unlabeled data

Question 7

What are examples of unsupervised learning?

Answer 7

Market basket analysis, customer segmentation

Question 8

What is reinforcement learning?

Answer 8

Learning to make sequences of decisions to maximize a reward

Question 9

What is a data set in ML?

Answer 9

A collection of information or records about a specific subject or entity

Question 10

What are attributes in a data set?

Answer 10

Features or variables that describe each record

Question 11

What is qualitative data?

Answer 11

Categorical data describing qualities (e.g., gender, grade)

Question 12

What are types of qualitative data?

Answer 12

Nominal (no order) and Ordinal (with order)

Question 13

What is quantitative data?

Answer 13

Numerical data that can be measured

Question 14

What is interval data?

Answer 14

Numeric data with known intervals but no true zero (e.g., temperature)

Question 15

What is ratio data?

Answer 15

Numeric data with a true zero (e.g., height, weight)

Question 16

What are the main steps in building a learning system?

Answer 16

Task understanding, data collection, data preparation, modeling, evaluation, deployment

Question 17

What is classification?

Answer 17

Assigning a label to input data based on learned patterns

Question 18

What is regression?

Answer 18

Predicting a numerical value based on input data

Question 19

What programming language is most used in ML?

Answer 19

Python

Question 20

What is a common library in Python for ML?

Answer 20

Scikit-learn

Question 21

What is the first step in problem-solving?

Answer 21

Problem formulation

Question 22

Name the main components in problem formulation.

Answer 22

Initial state, actions, transition model, goal test, path cost

Question 23

What is BFS (Breadth-First Search)?

Answer 23

Search algorithm that explores all nodes at the present depth before moving deeper

Question 24

What is DFS (Depth-First Search)?

Answer 24

Search algorithm that explores as far as possible along a branch before backtracking

Question 25

What is A* Search?

Answer 25

Heuristic search algorithm that balances cost and heuristic estimates

Question 26

What is the Vacuum World problem?

Answer 26

An AI agent cleans two squares using actions: Left, Right, Suck

Question 27

What does k in k-NN stand for?

Answer 27

The number of nearest neighbors considered

Question 28

Is k-NN a lazy learner?

Answer 28

Yes, it defers learning until a query is made

Question 29

Steps in k-NN classification?

Answer 29

Choose k, calculate distances, select k nearest neighbors, vote

Question 30

What distance metric is commonly used in k-NN?

Answer 30

Euclidean distance

Question 31

What is the formula for Euclidean distance in 2D?

Answer 31

√((x2 - x1)^2 + (y2 - y1)^2)

Question 32

Why choose an odd value for k in binary classification?

Answer 32

To avoid ties when voting

Question 33

What is the impact of irrelevant attributes in k-NN?

Answer 33

They can distort distance calculations and reduce accuracy

Question 34

What does normalization do in k-NN?

Answer 34

Rescales features to the same range, improving distance comparisons

Question 35

How do you normalize an attribute?

Answer 35

(x - min) / (max - min)

Question 36

In k-NN, what is the effect of different attribute scales?

Answer 36

Attributes with larger scales can dominate distance calculations

Question 37

In k-NN, what is the best practice with continuous attributes?

Answer 37

Use a consistent and appropriate distance metric like Euclidean

Question 38

Python code to create 3-NN classifier in sklearn?

Answer 38

knn = neighbors.KNeighborsClassifier(n_neighbors=3)

Question 39

Python code to train k-NN classifier?

Answer 39

fit = knn.fit(iris_train_ftrs, iris_train_tgt)

Question 40

Python code to evaluate accuracy?

Answer 40

metrics.accuracy_score(iris_test_tgt, preds)