CS3491- AIML

CAT2- Important Questions

2. What is ‘Over fitting’ and ‘Under fitting’ in Machine learning?

• Overfitting is when a machine learning model learns the training data too well,
including its noise and details, leading to poor performance on new, unseen data.

• Underfitting is when a model is too simple to learn the underlying pattern of the data,
resulting in poor performance on both the training and test data.

3. When would you use random forests and why?

Part B
1. Explain in details about Logistic Regression with an example. (8 Marks)
2. Discuss the Bayesian Linear Regression in detail. (8 Marks)
3. What is a classification tree? Explain the steps to construct a classification tree. List and
explain about the different procedures used. (16 Marks)
4. Summarize the working behavior of Support Vector Machines along with
Maximum margin Classifier with diagrams. (16 Marks)
5. Show the concept of decision tree classification algorithm in detail with example. (16
Marks)

Unit 4
Part A
1. Mention the benefit of voting.
 2. Define Stacking.
3. Recall the usage of Unsupervised learning.
4. Which Algorithm is called as Lazy learners? Why?
5. Write the three types of ensemble learning.
6. Define Ensemble Learning
7. How to Select Optimal K Value in K-Mean?
8. Explain Why K-NN is called as lazy learners.

Part B
1. Explain various learning techniques involved in Unsupervised Learning. (8
Marks)
2. Explain the concept of Gaussian Mixture Models (GMM). Describe the steps
involved in fitting a GMM using the Expectation-Maximization (EM) algorithm. (8
Marks)
3. Explain the various ensemble learning techniques in detail with neat diagram and
explanation. (16 Marks)
4. Predict the cluster using K-Mean Algorithm and explain the Steps involved for the
following set of four objects into two clusters A (3, 5 ), B (4,5), C (1,3), D (2,4). Consider
the Objects A and C as Initial Cluster Centers. (16 Marks)
5. Apply the K-Nearest Neighbors (KNN) algorithm to classify a new data point
using a suitable dataset. Demonstrate the working of the algorithm with a step-
by-step explanation of distance calculation, neighbor selection, and class
prediction. (16 Marks)
6. State the structure briefly about unsupervised learning (8 Marks)
7. List the application of clustering and identify advantages and disadvantages of
clustering algorithms. (8 Marks)
8. Explain in detail about combining multiple classifiers by Voting. (8 Marks)
9. Outline the steps in the AdaBoost algorithm with an example. (8 Marks)
10. Predict the cluster using K-Mean Algorithm and explain the Steps involved for the
following set of four objects into two clusters A (3,5), B (4,5), C (1,3), D (2,4). Consider
the Objects A and C as Initial Cluster Centers. (16 Marks)
11. Describe the working of the K-Nearest Neighbors (KNN) algorithm and apply it to
classify a new data point using a suitable example dataset. Explain the steps involved in
calculating distance, selecting neighbors, and predicting the output. (16 Marks)

Unit 5
Part A
1. How does error back propagation help in optimizing neural network
performance?
2. Differentiate computer and human brain
3. Explain the concept of a Multilayer Perceptron (MLP).
4. What is activation function and Mention its Types?
5. Compare whether stochastic gradient descent is same as gradient descent.
6. Why is ReLU better than softmax? Give the equation for both.
7. List the Components of perceptron.
 8. Differentiate computer and human brain

Part B
1. Explain in detail about Perceptrons and its types with neat diagram? (16 Marks)
2. Explain about ReLu, Hyperparameter tuning, Normalization, Regularization, Dropout?
(16 Marks)
3. Consider the perceptron having the weight w1=0.9 and w2=0.9 and α =
0.5(activation threshold) for AND Boolean function. Design and implement a
small SLP for AND logic gate. (16 Marks)
4. Build a model and Explain the algorithm for Back Propagation. (16 Marks)
5. How do you tune hypermaters for better neural network performance? Explain in detail.
(16 Marks)
6. Explain a deep feed forward network with a neat sketch (8 Marks)
7. Explain detail about activation functions? (8 Marks)
8. Consider the perceptron having the weight w1=0.9 and w2=0.9 and α =
0.5(activation threshold) for AND Boolean function. Design and implement a
small SLP for AND logic gate. (16 Marks)
9. Design and implement a simple neural network using the backpropagation algorithm.
Clearly describe each step involved in forward and backward pass. (16 Marks)