Machine Learning
Home My courses Previous Years 2021 Autumn Semester (2021-22)
Machine Learning Autumn 2021 Topic 19 Final Quiz
Started on Thursday, 18 November 2021, 12:10 PM
State Finished
Completed on Thursday, 18 November 2021, 1:10 PM
Time taken 59 mins 57 secs
Grade 64.00 out of 80.00 (80%)
Question 1 Generally, an ensemble method works better, if the individual base models have
Complete .............................?
Mark 4.00 out of
Select one:
4.00
a. Less correlation among predictions
Flag question
b. High correlation among predictions
c. Correlation does not have any impact on ensemble output
d. None of the above
Question 2 Which of the following statements are true for an Artificial Neural Network?
Complete
Select one or more:
Mark 6.00 out of
6.00 a. A single neuron with multiple inputs and the sigmoid activation function
may compute
Flag question
posterior of a class for a two-class classification problem in a
multidimensional feature
space.
b. A single neuron with two inputs and the signum activation function can
implement an OR
gate.
c. A single neuron with multiple inputs and the signum activation function
acts as a linear
discriminator in a multidimensional feature space for two linearly separable
classes.
d. A single neuron with two inputs and the signum activation function can
implement an
XOR gate.
Question 3 Which of the following points belong to the set of support vectors for an SVM
Complete classifying
Mark 4.00 out of them. CLASS 1: (1,1),(2,2),(4,1),(4,2)
6.00 CLASS 2: (2,-2), (3,-1), (3,-3), (4,-2)
Flag question
Select one or more:
a. (4,-2)
b. (3,-3)
c. (1,1)
d. (2,-2)
Question 4 Which of the following statements are NOT true for principal component analysis
Complete (PCA)?
Mark 4.00 out of
Select one or more:
6.00
a. It provides a transformation of points preserving distances among them in
Flag question
the transformed space.
b. It provides a direction of projection of a feature vector for optimum
separation of data
points belonging to two classes.
c. It computes minimum dimensional space sufficient to represent the data.
d. It computes a component along the direction providing maximum variance
among all
possible projections.
Question 5
Complete
Mark 4.00 out of
4.00
Flag question
What is the sum of eigen values of the covariance matrix.
Answer upto 2 decimal places
Answer: 12.19
Question 6 Consider the below statements about the kernel in SVM. Which of them are true?
Complete
Mark 6.00 out of
Select one or more:
6.00 a. A kernel function explicitly maps low dimensional points to high
dimensional space.
Flag question
b. Given the feature space X, a kernel function f may be mathematically
denoted by f : X × X → R, where R is the set of real numbers.
c. Any distance function satisfying properties of a metric can be used in
defining a kernel
function.
d. A kernel function may provide negative values.
Question 7 Consider a linear discriminant function g(x1 , x2 , x3 , x4 ) = 4x1 − 5x2 + 3x3 + 8x4
Complete − 5 for
Mark 6.00 out of classifying a feature vector X = (x1 , x2 , x3 , x4 ). Which of the following
6.00 statements are true
Flag question with respect to this classifier?
Select one or more:
a. The class labels of the NULL feature vector (i.e. x1 = x2 = x3 = x4 = 0) and
(1, 1, 1, 1) are
different.
b. If g(x) classifies a set of labeled training examples without any error, there
does not exist
any other linear discriminant function with the same performance.
c. If a feature point lies on g(x), its class label is ambiguous.
d. The class boundary is a hyperplane in a 4-D real space whose normal is (8,
−10, 6, 16)
Question 8
Complete
Mark 0.00 out of
8.00
Flag question
Answer upto 4 decimal places
Answer: -0.024
Question 9 While exploring a nondeterministic environment, an agent maintains running
Complete averages of
Mark 6.00 out of reward values of states for every action. Consider the agent takes an action a at
6.00 state s to
Flag question move at state q and obtains a reward 50. The present reward value of the state
action pair
(s, a) is 100, and the maximum possible discounted reward at q is estimated at
80. Given the
learning rate 0.3, what would be the updated reward value at (s, a)?
Answer: 109
Question 10 Consider a multilayer feed forward network which takes a 12-D feature vector as
Complete input and
Mark 6.00 out of produces a 3-D output vector. The network has two hidden layers with number of
6.00 neurons 13
Flag question and 5 for the first and second hidden layers, respectively. Please note that the
input vector is
the input to the first hidden layer. What is the dimension of the parametric space
over which
the optimization process would be carried out to train the network?
Answer: 257
Question 11 While training a Adaboost algorithm on a binary classification dataset, the first
Complete base learner incurs a 30% error on the training data. What will the weight of the
base learner in the final prediction of the ensemble classifier?
Mark 4.00 out of
4.00
Flag question answer upto 4 decimal places.
Answer: 0.8473
Question 12 Given set of five features {x1,x2,x3,x4,x5}. Consider Sequential Forward Selection
Complete method.
Mark 4.00 out of We have selected S 1 = {x1, x2} till now. Let the validation error on a selected set
4.00 of features
Flag question S be denoted by E(S ). Consider below information regarding validation set error
and answer
which feature should be added next.
E(S 1) = 0.2
E(S 1 ∪ {x3}) = 0.25
E(S 1 ∪ {x4}) = 0.30
E(S 1 ∪ {x5}) = 0.35
Select one:
a. x4
b. x5
c. x3
d. None
Question 13 Which of the following can be true for selecting base learners for an ensemble?
Complete
Mark 0.00 out of 1. Different learners can come from same algorithm with different hyper
4.00 parameters.
Flag question 2. Different learners can come from different algorithms.
3. Different learners can come from different sets of training samples.
4. Different learners can perform different tasks.
Select one:
a. 1 and 2
b. 1, 2, 3 and 4
c. 1, 2 and 3
d. only 1
Question 14 If each base-learner is an i.i.d. and correct with probability 0.75, what is the
Complete probability that a majority vote over 5 classifiers gives the correct answer?
Mark 4.00 out of
4.00 answer upto 5 decimal places.
Flag question
Answer: 0.89648
Question 15 Which of the following statements are true for a sigmoid activation function?
Complete
Select one or more:
Mark 6.00 out of
6.00
a. Its functional value can solely provide the value of logit function at that
point.
Flag question
b. Its functional value can solely provide the value of its differentiation at that
point.
c. It is not differentiable at every point.
d. It is not defined at every point.
Finish review
QUIZ NAVIGATION
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Show one page at a time
Finish review
You are logged in as Rajat Bachhawat Log
out
