Assignment Week 5
Assignment Week 5
1. What would be the ideal complexity of the curve which can be used for separating the two
classes shown in the image below?
A) Linear
B) Quadratic
C) Cubic
D) insufficient data to draw conclusion
Answer: A
(The blue point in the red region is an outlier (most likely noise). The rest of the data is
linearly separable.)
Answer: A
Linear regression error values have to be normally distributed only.
3. Which of the following methods do we use to best fit the data in Logistic Regression?
A) Manhattan distance
B) Maximum Likelihood
C) Jaccard distance
D) Both A and B
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
Answer: B
In logistic regression, maximum likelihood is used as estimation methods for fitting the data.
4. Imagine, you have given the below graph of logistic regression which shows the relationships
between cost function and number of iterations for 3 different learning rate values (different
colors are showing different curves at different learning rates).
Suppose, you save the graph for future reference but you forgot to save the value of different
learning rates for this graph. Now, you want to find out the relation between the leaning rate
values of these curve. Which of the following will be the true relation?
Note: 1. The learning rate for blue is L1.
2. The learning rate for red is L2.
3. The learning rate for green is L3.
A) L1>L2>L3
B) L1=L2=L3
C) L1<L2<L3
D) None of these
Answer: C
If you have low learning rate means your cost function will decrease slowly but in case of
large learning rate cost function will decrease very fast.
Answer: A
This is true because the support vectors only affect the boundary.
6. Suppose you are dealing with 3 class classification problem and you want to train a SVM
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
model on the data for that you are using One-vs-all method.
How many times we need to train our SVM model in such case?
A) 1
B) 2
C) 3
D) 4
Answer: C
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
In a N-class classification problem, we have to train the SVM at least N times in a one vs
all method.
A) 1
B) 2
C) 1 and 2
D) None of these.
Answer: C
Kernels are used in SVMs to map low dimensional data into high dimensional feature
space to classify non-linearly separable data. It is a similarity function between low-
dimensional data points and its high dimensional feature space to find out what data points
can be mapped into what sort of feature space.
8. Suppose you are using RBF kernel in SVM with high Gamma value. What does this signify?
A) The model would consider even far away points from hyperplane for modelling.
B) The model would consider only the points close to the hyperplane for
modelling.
C) The model would not be affected by distance of points from hyperplane for
modelling.
D) None of the above
Answer: B
The gamma parameter in SVM tuning signifies the influence of points either near or far
away from the hyperplane.
For a low gamma, the model will be too constrained and include all points of the training
dataset, without really capturing the shape.
For a higher gamma, the model will capture the shape of the dataset well.
9. Below are the labelled instances of 2 classes and hand drawn decision boundaries for
logistic regression. Which of the following figure demonstrates overfitting of the training data?
A) A
B) B
C) C
D) None of these
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
Answer: C
In figure 3, the decision boundary is very complex and unlikely to generalize the data.
10. What do you conclude after seeing the visualization in previous question?
C1. The training error in first plot is higher as compared to the second and third plot.
C2. The best model for this regression problem is the last (third) plot because it has
minimum training error (zero).
C3. Out of the 3 models, the second model is expected to perform best on unseen data.
C4. All will perform similarly because we have not seen the test data.
A) C1 and C2
B) C1 and C3
C) C2 and C3
D) C4
Answer: B
From the visualization, it is clear that the misclassified samples are more in the plot A when
compared to B. So, C1 is correct. In figure 3, the training error is less due to complex
boundary. So, it is unlikely to generalize the data well. Therefore, option C2 is wrong.
The first model is very simple and underfits the training data. The third model is very
complex and overfits the training data. The second model compared to these models has
less training error and likely to perform well on unseen data. So, C3 is correct.
We can estimate the performance of the model on unseen data by observing the nature of
the decision boundary. Therefore, C4 is incorrect
End