Machine Learning for Engineering and Science A... [Link]
mucharladonald@[Link] ▼
Courses » Machine Learning for Engineering and Science Applications
Announcements Course Ask a Question Progress FAQ
Unit 7 - Week 4
Register for
Certification exam Assignment 4
The due date for submitting this assignment has passed. Due on 2019-02-27, 23:59 IST.
Course
outline Assignment submitted on 2019-02-27, 21:53 IST
PLUS
How to access the 1) 1 point
portal
Matlab and
Learning Modules
Pre-Requisite
assignment
Week 1
Week 2
Week 3
Week 4 Yes, the answer is correct.
Score: 1
The Learning
Paradigm Accepted Answers:
A Linear
Regression
Example
Linear 2) 1 point
Regression
Least Squares
Gradient
Descent
Coding Linear
Regression
Generalized © 2014 NPTEL - Privacy & Terms - Honor Code - FAQs -
Function for
A project of
Linear In association with
Regression
Goodness of Fit
Funded by
Bias-Variance
1 of 5 4/24/19, 2:45 PM
Machine Learning for Engineering and Science A... [Link]
Trade Off Powered by
Gradient
Descent
Algorithms
Quiz :
Assignment 4
Additional
Materials
Week - 4
Feedback Form
Week 5
Week 6
Week 7
Week 8
Week 9
PLUS
Week 10 Yes, the answer is correct.
Score: 1
Week 11 Accepted Answers:
Week 12
Download Videos
TEXT
TRANSCRIPTS
Interaction
session 3) Which of the following statements are True? Check all that apply: 1 point
If a learning algorithm is suffering from high bias, only adding more training examples may
not improve the test error significantly.
A model with more parameters is more prone to overfitting and typically has a higher
variance.
When debugging learning algorithms, it is useful to plot a learning curve to understand if
there is a high bias or high variance problem.
Increasing degree of the polynomial in curve fitting will increase the bias in the model
Yes, the answer is correct.
Score: 1
Accepted Answers:
If a learning algorithm is suffering from high bias, only adding more training examples may not
improve the test error significantly.
A model with more parameters is more prone to overfitting and typically has a higher variance.
When debugging learning algorithms, it is useful to plot a learning curve to understand if there is a
high bias or high variance problem.
4) The figure below shows the plot of the learning curves of a learning algorithm. It is found 1 point
that it has an unacceptably high error on the test set. What is the algorithm suffering?
2 of 5 4/24/19, 2:45 PM
Machine Learning for Engineering and Science A... [Link]
High Variance
High Bias
High Variance and Low bias
PLUS
None
Yes, the answer is correct.
Score: 1
Accepted Answers:
High Bias
5) Suppose you have implemented a regularized linear regression model. You observe that on 1 point
the held out testing set, the model makes unacceptably large errors with its predictions. However, you observe
that the model performs well (has a low error) on the training set. Which of the following steps can be
incorporated to lower the error on testing dataset. Select all that apply.
Try using a smaller set of the features
Try decreasing the regularization parameter λ
Get more training examples
Use fewer training examples
No, the answer is incorrect.
Score: 0
Accepted Answers:
Try using a smaller set of the features
Get more training examples
6) Suppose you have implemented a regularized linear regression model. You observe that on the 1 point
held out testing set, the model makes unacceptably large errors with its predictions. Furthermore, you observe
that the model performs poorly on the training set. Which of the following steps can be incorporated to lower
the error on the testing dataset. Select
all that apply
Try to obtain an additional set of features
Try increasing the regularization parameter λ
Get more training examples
3 of 5 4/24/19, 2:45 PM
Machine Learning for Engineering and Science A... [Link]
Try adding polynomial features
Yes, the answer is correct.
Score: 1
Accepted Answers:
Try to obtain an additional set of features
Get more training examples
7) Suppose you are training a regularized linear regression model. Check which of the following 1 point
statements are true? Select all that apply.
The regularization parameter λ value is chosen so as to give the lowest training set error
The regularization parameter λ value is chosen so as to give the lowest cross validation error
The regularization parameter λ value is chosen so as to give the lowest test set error
The performance of a learning algorithm on the training set will typically be better than its
performance on the test set
Yes, the answer is correct.
Score: 1
PLUS
Accepted Answers:
The regularization parameter λ value is chosen so as to give the lowest cross validation error
The performance of a learning algorithm on the training set will typically be better than its performance on the
test set
8) 1 point
1.03125
2.03125
3.03125
4.03125
No, the answer is incorrect.
Score: 0
Accepted Answers:
2.03125
9) 1 point
4 of 5 4/24/19, 2:45 PM
Machine Learning for Engineering and Science A... [Link]
Yes, the answer is correct.
Score: 1
Accepted Answers:
10)What is the cost now? 1 point
0.4292
PLUS
1.4292
2.4292
3.4292
Yes, the answer is correct.
Score: 1
Accepted Answers:
0.4292
Previous Page End
5 of 5 4/24/19, 2:45 PM