ML Deep Learning Fundamentals Applications Week 3 Answers

Are you searching for reliable ML Deep Learning Fundamentals Applications Week 3 Answers 2024? Look no further! Our solutions are designed to provide clear, detailed answers, helping you navigate your NPTEL course with confidence.

ML Deep Learning Fundamentals Applications Week 3 Answers (July-Dec 2024)

Q1.The bandwidth parameter in the Parzen Window method determines:
The number of neighbors to consider for classification
The size of the neighborhood around a test instance
The dimensionality of the feature space.
The complexity of the classifier

Q2.If the number of data samples becomes very large.
Bayesian Estimation is worse than MLE
Maximum Likelihood estimates are slightly bad
Bayesian Estimation performs same as MLE
None

These are ML Deep Learning Fundamentals Applications Week 3 Answers

Q3. What happens when k=1 in k -Nearest Neighbor algorithm:
Underfitting
Overfitting
High testing accuracy
All the above

Q4. There are 18 points in an axis plane namely –
[(0.8,0.8)t,(1,1)t,(1.2,0.8)t,(0.8,1.2)t,(1.2,1.2)t]
,
belong to class 1;
[(4,3)t,(3.8,2.8)t,(4.2,2.8)t,(3.8,3.2)t(4.2,3.2)t,(4.4,2.8)t,(4.4,4.4)t]
,
belong to class 2;
[(3.2,0.4)t,(3.2,0.7)t,(3.8,0.5)t,(3.5,1)t,(4,1)t,(4,0.7)t]
,
belong to class 3.
A new point
P=(4.2,1.8)t
introduces into the map. The point P belongs to which class? Use k
-nearest neighbor technique with k=5
to calculate the result.

Class 1
Class 2
Class 3
None of the above

These are ML Deep Learning Fundamentals Applications Week 3 Answers

Q5. Suppose we have two training data points located at 0.5 and 0.7, and we use 0.3 as its rectangle window width. Using the Parzen window technique, what would be the probability density if we assume the query point is 0.5?
0.5
0.75
2.22
1.67

Q6. Suppose that X is a discrete random variable with the following probability
mass function: where is a parameter.
(0≤θ≤1)

The following 10 independent observations were taken from such a distribution:
(3,0,2,1,3,2,1,0,2,1)
. What is the maximum likelihood estimate of θ
?
2
1
0.5
0

Q7. Which of the following statements are true about k

nearest neighbor (KNN)-
Odd value of “K” preferred over even values.
Does more computation on test time rather than train time.
Work well with high dimension.
The optimum value of K for KNN is highly independent on the data.

Q8.The disadvantage of using k-NN as classifier:
Fails while handling large dataset
Fails while handling small dataset
Sensitive to outliers
Training is required

These are ML Deep Learning Fundamentals Applications Week 3 Answers

Q9. Consider single observation X that depends on a random parameter .Suppose θ
has a prior distribution

fθ(θ)=λe−λθforθ≥0,λ>0fxθ(x)=θe−θx|x|>0

Find the MAP estimation of θ

1λ+X
1λ−X
λX
X

Q10.The MLE for the data samples X={x1,x2,…,xi,…,xk} with the Bernoulli distribution is

n⋅xk

xkn

Mean of xi
None