D DC203 Pages: 2
Page 1 of 2
Reg
APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
SECON SEMESTER (Second Year Direct)
FOURTH SEMESTER MCA (Regular) DEGREE EXAMINATION, APRIL 2018
Course Code: RLMCA208
Course Name: INTRODUCTION TO MACHINE LEARNING
Max. Marks: 60 Duration: 3 Hours
PART A
Answer all questions, each carries 3 marks. Marks
1 What is the purpose of Ordinary Least Square Estimation?
2 Give a sample scenario where decision tree can be used for classifying data?
3 Explain the structure of a single artificial neuron with a diagram.
4 What is deep learning?
5 Give one method to choose a maximum margin hyperplane for SVM classifiers?
6 What is a Support Vector?
7 What are the advantages of K-fold cross validation?
8 How Boosting process improves model performance?
PART B
Each question carries 6 marks.
9 Explain PCA and its steps in detail.
OR

Describe any 6 different measurements of central tendency spread with
relevant examples?

10
We have data from survey and objective testing with two attributes (acid
durability and strength) to classify whether a special tissue is good or not.
Here is four training samples.
X1(Acid Durability) X2(Strength) Y(Class)
7 7 BAD
7 4 BAD
3 4 GOOD
1 4 GOOD
Now the factory produces a new tissue paper that pass the test with x value

D DC203 Pages: 2
Page 2 of 2
3 and Y value 5. Find the classification of this new paper from the data of
trained samples using KNN algorithm.
OR

Write a note on Bayes theorem and illustrate the method for predicting
probabilities with an example.

11
Differentiate Simple Linear Regression Multiple linear regression with an
example.

OR

Explain the divide and conquer approach for the construction of decision
trees with an example.

12 Explain any 3 characteristics of neural networks?
OR
How does a Perceptron learn the appropriate weights using delta rule?
13 How SVM handles non- linearly separable data.
OR

How Classification using hyper planes is possible? What is Maximum Margin
Hyperplane?

14
How will you evaluate the performance of a model using confusion
matrices? Justify answer using the statistics Accuracy, Precision and
Recall.

OR

How ensembles learning improve model performance? Explain any two
ensemble based methods.