SRM VALLIAMMAI ENGINEERING COLLEGE

(An Autonomous Institution)

SRM Nagar, Kattankulathur – 603 203

DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING

QUESTION BANK

VI SEMESTER

1904012 MACHINE LEARNING

Academic Year 2021 – 22 EVEN

Prepared by

Mr.A.PANDIAN., Assistant Professor (Sr.G)

Ms.V.Venmathi., Assistant Professor (O.G)
 Unit -I
Machine learning: What and why? - Examples of Machine Learning Applications - Types Of Machine
Learning Supervised Learning - Machine Learning Process- The Curse of Dimensionality, Over fitting -
Training, Testing, and Validation Sets-The Confusion Matrix & Basic Statistics-Bias-Variance Tradeoff.
.
PART – A
Q.No Questions Competence BT Level

1 What is Machine learning? What is the need of it? Remember BTL1

2 List applications of machine learning. Analyze BTL 4
3 Distinguish between supervised and unsupervised learning. Understand BTL 2
4 Define activation function. Remember BTL 1
5 Summarize the steps involved in ML process. Understand BTL 2
6 Compute the accuracy metrics in confusion matrix. Apply BTL3
7 Write short notes on issues in machine learning. Analyze BTL 4
8 Define the generalization in ML. Evaluate BTL 5
9 State about Supervised Learning. Remember BTL 1
10 Depict the important objectives of machine learning? Evaluate BTL 5
11 Label the term Over fitting. Remember BTL 1
12 Illustrate an example for Data sparsity. Create BTL 6
13 Outline the confusion matrix. Remember BTL 1
14 Classify the applications of Machine Learning. Remember BTL1
15 Examine the process of Machine learning. Apply BTL 3
16 Mention with examples some useful applications of machine Understand BTL 2
learning.
17 How would you use the Curse of Dimensionality in ML. Apply BTL 3
18 Sketchconfusion matrix with an example. Create BTL 6
19 Compare the Precision and Recall. Analyze BTL 4
20 Mention the term variance and covariance. Understand BTL 2
PART-B (13 MARK )
1 Examine in detail ofmachine learning process. (13) Analyze BTL 4
2 Explain the Supervised learningwith a Classification problem.(13) Understand BTL 2

3 Justify the Issues in Machine Learning and Curse of Dimensionality.(13) Evaluate BTL 5

4 Write down the concept of Overfitting,Training , Testing and Validation Remember BTL 1
sets with examples. (13)
5 Sketch and explain the confusion Matrix with an example. (13) Apply BTL 3
6 Illustrate with diagrams of Machine Learning Process with an Apply BTL 3
example.(13)
7 Develop an example for Naïve Bayes Classifier and discuss in Create BTL 6
detail.(13)
8 (i) Write short note on the statistics involved in Machine Learning.(7) Remember BTL 1
 (ii) Define Bias Variance Tradeoff with example. (6)
9 Explain the following: Understand BTL 2
(i) Confusion Matrix. (7)
(ii)concept of Precision and Recall with example.(6)
10 What is Training, Testing and Validation sets,explain with Remember BTL 1
examples.(13)
11 State the importance of Naïve Bayes Classification and explain the Bias Remember BTL 1
and Variance. (13)
12 Summarize in detail about types of Regression and Classification in Understand BTL 2
Supervised Learning. (13)
13 (i) Write in detail about Supervised Learning. (7) Analyze BTL 4
(ii)Discuss the various application of Machine Learning. (6)

14 Compare the Regression and Classification with an example. (13) Analyze BTL 4
PART - C (15 MARK)
1 Given the confusion matrix [[ 737] [4 144]],Develop and Calculate Create BTL 6
Accuracy, Precision, Recall and Specificity, Explainindetail.(15)
2 Explain the following for designing a learning system Analyze BTL 4
(i)choosing a function approximation algorithm.(5)
(ii)estimating training values.(4),adjusting theweights.(4)
(iii) the finaldesign.(2)
3 Describe in detail, how Machine Learning Algorithms can be tested Remember BTL 1
withnecessaryexamples. (15)
4 Summarize the Learning Process & Explain Supervised Learning in Evaluate BTL 1
detail, Give Necessary examples. (15)
 Unit -II

Hebb’sRule-NeuralNetworks-ThePerceptron-LinearSeparability&LinearRegression.TheMulti-layer
Perceptron: Biases, Algorithm - Local minima and Stochastic gradient Descent Examples Of Using The
MLP : Regression Problem & Classification Example - DerivingBack-Propagation
PART - A

Q.No Questions Competence BT Level

1 Write down the significance of Back propagation Algorithm. Remember BTL4

2 What do you mean by Gradient Descent? Remember BTL1
3 List out different output activation functions. Remember BTL1
4 Mention the feature Local minima. Understand BTL2
5 State the Hebb’s Rule. Remember BTL1
6 Compare the difference between local and distributed representation. Apply BTL2
7 List the advantage of Linear Separability & Linear Regression. Remember BTL1
8 Name the two parts of MLP. Understand BTL2
9 Summarize the significance Regression problem. Understand BTL2
10 Write the error function used for the perceptron. Remember BTL1
11 Examine the Neural Network and its functionality. Apply BTL 3
12 How will you distinguish the deriving Back Propagation. Apply BTL 3
13 Outline the features ofPerceptron. Apply BTL 3
14 Depict the principles of gradient descent. Evaluate BTL 5
15 Define Biases. Remember BTL 1
16 Assess the characteristics optimal separating . Analyze BTL 4
17 Interpret the importance Local minima. Analyze BTL 4
18 Justify the need of Back propagation of error in ML. Evaluate BTL‐5
19 Sketch block diagram of McCulloch. Create BTL‐6
20 How perecptron learning algorithm used in CNN Create BTL‐6
PART-B (13 MARK )
Explain how to learn Multilayer Networks using Backpropagation
1 Algorithm. (13) Remember BTL 1
Mention the steps of Back propagation algorithm. (7)
2 Remember BTL 1
Write down the training rule for output unit weights (6)
Write short note on the following
3 Remember BTL 1
(i) Multilayer perceptron algorithm. (7)
(ii) MLP is designed to be a batch algorithm (6)
(i) What is Local Minima and explain? (7)
4 Remember BTL 1
(ii) Discuss in detail about picking up Momentum (6)
Explain the following
(i) Data preparation in MLP (5) (ii)Amount of training Data (4)
5 Understand BTL 2
(iii)Number of Hidden layers (4)
6 (i)Describe the Linear separability and Linear Regression.(7) Understand BTL 2
 (ii)How to train test and validate the MLP. (6)
(ii) Summarize the Regression Problem of using MLP. (7)
7 (ii) How Data compression is performed (6) Understand BTL 2
Illustrate Generalization of multivariate data and explain in detail
8 about Classification Problem.(13) Apply BTL 3
9 Sketch andExplain in detail about Neural network. (13) Apply BTL 3
(i) Write in detail about MLP with example.(7)
10 (ii) Investigate the steps to derive Back Propagation Network (6) Analyze BTL 4
Compare thePerceptron Convergence Theorem and Linear
11 Regression with example (13). Analyze BTL 4
12 Examine in detail of MccCulloch and Pits Neuronal Model. (13) Analyze BTL 4
With necessary diagramExplain how the weights of the network are
13 trained (13) Evaluate BTL 5
14 Justify your answer why we use Interpolation andBasis function.(7) Create
Outline thedetail about MLP Algorithm.(6) BTL 6
PART - C (15 MARK )
1 Write about the MLP as a Universal Approximator. (15) Evaluate BTL5
(i) Develop an example and explain about Linear Regression. (8)
2 (ii)Describe the Linear Separability. Create BTL6
(7)
Illustrate some examples of using MLP and the four types of
3 problems that are generally solved using MLP. Evaluate BTL5
(15)
Elaborate following in detail :
4 (i) taskinvolved.(5)
(ii) Inputchoice.(5) Create BTL 6
(iii) Input encoding and output encoding.(5)
 Unit -III
Linear Discriminant Analysis (LDA) - PRINCIPAL COMPONENTS ANALYSIS (PCA), Factor Analysis -
IndependentComponentsAnalysis-PROBABILISTICMODEL-GaussianMixtureModels:EMAlgorithm
- Nearest Neighbour Methods - Support Vector Machines
PART - A
Q.No
Questions Competence BT Level

1 Define LDA. Remember BTL1

2 What is Support Vector Machine? Remember BTL1
3 Sketch thePrincipal Component Analysis. Remember BTL1
4 Draw an example for Gaussian Mixture Models. Apply BTL3
5 Outline about Factor Analysis. Understand BTL2
6 List out the advantages of SVM. Remember BTL1
7 Distinguish between LDA and PCA. Apply BTL3
8 State the basic principle of EM algorithm. Remember BTL1
9 Outline the significance Nearest Neighbour methods. Understand BTL2
10 Classify the different ways for Dimensionality Reduction. Apply BTL3
11 Point out the distance measure in Nearest neighbor methods. Analyze BTL4
12 Compare between PCA and its relation with Multi Layer Analyze BTL4
Perceptron.
13 Summarize theLocally Linear embedding. Evaluate BTL5
14 Outline the Independent Component analysis. Remember BTL1
15 Enumerate theKernel PCA. Understand BTL2
16 Write the algorithmic steps in PCA. Evaluate BTL5
17 Justify the efficient distance computation methods. Create BTL6
18 write the distance measure used in probabilistic model. Create BTL6
19 Mention the featuresSVM Regression? Understand BTL2
20 State the importance General Expectation-Maximisation (EM) Analyze BTL4
Algorithm.
PART-B (13 MARK )
(i) What is Principal Component Analysis.(7)
1 Remember BTL1
(ii) Write the algorithm of PCA.(6)
(i)Describe the PCA relation with MLP. (7)
2 Remember BTL1
(ii)Explain Kernel PCA Algorithm. (6)
3 (i) Outline the Gaussian Mixture Models.(7) Analyze BTL4
(ii) How Independent Component Analysis is done.(6)
Discuss the Probabilistic Learning methods involved in Machine
4 Understand BTL2
Learning. (13)
Illustrate the Gaussian Expectation-Maximization Algorithm
5 Apply BTL3
with necessary example. (13)
6 Outline about the Nearest Neighbor methods withexample.(13) Remember BTL1
Interpret the need of Support Vector Machine with necessary
7 Apply BTL3
examples.(13)
 Write about Factor Analysisand Expectation-Maximization (Em)
8 Remember BTL1
Algorithm.(13)
Explain the following :
9 (i)Gaussian.(6) Analyze BTL4
(ii)EM Algorithm. (7)
(i) Mention the feature Nearest Neighbor Algorithm. (7)
10 Understand BTL2
(ii)Write in detail about Nearest Neighbor Smoothing.(6)
Summarize the basic idea of KD-Tree and Support Vectors.(13)
11 Understand BTL2
With necessary diagram to explain Independent Component
12 Analysis and Linear Discriminant Analysis. (13) Evaluate BTL5
13 Justify the need in ML use of optimal Separation in Support Analyze BTL4
Vector Machine. (13)
Illustrate with diagrams of the XOR and SVM Regression. (13)
14 Create BTL6
PART - C (15 MARK)
Illustrate EM Algorithm and explain the algorithm for thecoin
1 Create BTL6
flippingExample. (15)
Discuss in detail the working of Support Vector Machine, its
2 Evaluate BTL 5
implementationandexamples. (15)
(i) Explain in detail about GaussianMixture Model. (7)
3 Evaluate BTL 5
(ii) Write the General Expectation(EM)Algorithm. (8)
Justify the need of Dimensionality Reduction by explaining the
following methods
4 (i) LinearDiscriminant Analysis (5)
Create BTL6
(ii) PrincipalComponentAnalysis (5)
(iii) IndependentComponentAnalysis (5)
 Unit -IV
Evolutionary Learning - The Genetic Algorithms (GA)-Reinforcement Learning -Decision Trees -
CLASSIFICATIONANDREGRESSIONTREES(CART)-EnsembleLearning:Boosting,Bagging,Random
Forests - Unsupervised Learning : K-Means – Algorithm - VectorQuantisation.

PART - A

Q.No Questions Competence BT Level

1 Distinguish between crossover and mutation. Analyze BTL4

2 Define Genetic Algorithm. Remember BTL1
3 Write the concept involved in evaluating fitness. Understand BTL2
4 State the limitation of mutation in GA.. Remember BTL1
5 Summarize about the Genetic Algorithm. Understand BTL2
6 Point out the use of Vector Quantisation. Apply BTL3
7 Outline the disadvantage of Genetic Algorithm. Analyze BTL4
8 Justify the importanceBoosting and Bagging. Evaluate BTL5
9 Mention the about Markov property. Remember BTL1
10 What are the features of a Markov decisionProcess. Evaluate BTL5
11 How K-Means Algorithm is used in Unsupervised Learning. Remember BTL1
12 Compile and write the basic concept involved in CART. Create BTL6
13 Outline the Gini Impurity. Remember BTL1
14 Identify how reinforcement learning maps states to action. Remember BTL1
15 Mention the advantage Genetic Programming. Apply BTL3
16 List the advantage of reward function. Understand BTL2
17 Identify the principle of the entropy function. Apply BTL3
18 Differentiate Sarsa and Q-learning. Create BTL6
19 State the importance of ID3. Analyze BTL4
20 Outline the significance of the shortly Random Forest Algorithm. Understand BTL2
PART-B (13 MARK )
Write the use of Genetic Algorithm.(7)
1 Analyze BTL4
Discuss an example that illustrate the use of Genetic Algorithm.(6)
Summarize the about Reinforcement Learning and elaborate on Markov
2 Understand BTL2
Decision Process. (13).
Illustrate with necessary diagrams about Decision Tree. (7)
3 Evaluate BTL5
Write in detail Classification and Regression Tree(CART).(6)
4 Describe Random Forest Algorithm with an example. (13) Remember BTL1
Explain in detail about Boosting and Bagging.(13)
5 Apply BTL3
Outline the algorithm involved in Unsupervised Learning
6 Create BTL6
Environment. (13)
Briefly explain the Knapsack Problem using Genetic Algorithm.(13)
7 Remember BTL1
8 Write thedetail about Generating Offspring Genetic Operators. (13) Understand BTL2
 List the different between Sarsa and Q-learning, give examlple of the
9 Remember BTL1
reinforcement learning. (13)
Mention the Knapsack problem for GA and explain in detail. (13)
10 Remember BTL1
(i)Outline the working of reinforcement learning. (7)
11 Create BTL6
(ii)Draw the block diagram of reinforcement learning.(6)
Summarize theabout the ID3,how Classification is done with an
12 Create BTL 6
Example.(13)
Write short notes on
13 Remember BTL 1
(i) CART.(7)
(ii) Combine classifiers.(6)
Explain in detail about
14 (i) K-Means Clustering algorithm(7) Create BTL 6
(ii) VectorQuantisation.(6)
PART - C (15 MARK )
Elaborate the Decision Tree and explain the implementation of
1 Create BTL 6
Decision Tree (15)
Write in detail about Reinforcement Learning, Give an Example
2 Create BTL 6
implementation and write downitsuses. (15)
Choose two destination with different routes connecting them.Apply
3 Analyze BTL 4
genetic algorithm to find the optional path basedon distance. (15)
(i) Describe about K-Means Algorithm withanexample. (8)
4 Evaluate BTL 5
(ii) Explain about Normalization andVectorQuantization. (7)
 Unit -V
Bayesian Networks - Markov Random Fields - Hidden Markov Models (HMMS) - Markov Chain Monte Carlo
(MCMC) Methods - Deep Belief Networks (DBN)

PART - A

Q.No Questions Competence BT Level

1 Point the concept involved in Bayesian Networks. Analyze BTL4

2 Define MCMC. Remember BTL1
3 Distinguish random numbers and Gaussian Random numbers Understand BTL2
4 List the major function block of Deep Belief Network. Remember BTL1
5 Outline the advantages Markov Chains. Understand BTL2
6 Sketch the Graphical models. Apply BTL3
7 Write the Various MCMC methods. Analyze BTL4
8 Summarize thegraphical models. Evaluate BTL5
9 Outline line about HMM. Remember BTL1
10 Write Variable Elimination Algorithm. Evaluate BTL5
11 Mention the purpose of Gibbs Sampling. Remember BTL1
12 Distinguish between CNN, Bayesian Belief network. Create BTL6
13 Compare the Deep learning and Machine Learning. Analyze BTL4
14 Write about Baum–Welch or Forward–Backward Algorithm. Remember BTL1
15 Point out the various tracking methods. Apply BTL3
16 Draw the block diagram of Deep Learning. Understand BTL2
17 Mention the Markov Random Field Image Denoising Algorithm. Apply BTL3
18 State in detail about Viterbi algorithm. Create BTL6
19 Examine the Markov Random Fields. Analyze BTL4
20 Depict the Forward algorithm. Understand BTL2
PART-B (13 MARK )
1 Outline the concept involved in Markov Random Fields. (13) Analyze BTL4
(i) Discuss Gaussian Random numbers.(7)
2 Understand BTL2
(ii) Describe Markov Random Fields.(6)
(i) Write the Rejection Sampling Algorithm.(7)
3 Evaluate BTL5
(ii) Sketch the histogram of a mixture of two Gaussians.(6)
Mention the Sampling –importance resampling algorithm and Gibbs
4 Remember BTL1
Sampling.(13)
(i) Explain in detail about MCMC.(7)
5 Apply BTL3
(ii) Summarize the Simulated Annealing by a distribution.(6)
Structure two graphical models and show the various relationship
6 between the nodes.(13). Create BTL6
(i)Write short on variable elimination algorithm. (7)
7 Remember BTL1
(ii)Depict the Approximate Inference. (6)
Identify the features involved in Simulated Annealing and Gibbs
8 Understand BTL2
Sampling. (13).
9 State how the decoding problem occurs by viterbi algorithm. (13) Remember BTL1
(i) List out the Markov Random fields.(7)
10 Remember BTL1
(ii) Point out the Markov Random Field Image Denoising Algorithm.(6)
 (i)Write in detail forward algorithm. (7)
11 Create BTL6
(ii)Discuss HMM forward algorithm. (6)
12 Summarize the HMM Baum-Welch(forward –Backward )algorithm with Create BTL 6
example. (13)
(i) Write about Tracking methods.(7)
13 Remember BTL 1
(ii) Identify hidden Markov models in detail.(6)
Elaborate following
14 Create BTL 6
(i)Kalman Filter Algorithm. (7)
(i)Particle Filter. (6)
PART - C (15 MARK )
Discuss about Bayesian Network and discuss in detail with “Exam Fear”
1 Create BTL 6
example. Also explain the making ofBayesianNetworks. (15)
Write the concept involved in Deep Belief Network and its algorithm
2 also justify how Deep learning is the most prominent field in current Create BTL 6
generation. (15)
Explain about Markov Chain Monte Carlo Methods, algorithm and
3 Analyze BTL 4
implementationwithexample. (15)
Describe in detail about the Hidden Markov Models.(HMMS) also
4 Evaluate BTL 5
generate the Forward, Viterbi and BaumWelchAlgorithm. (15)