Scribd
Upload
31 views2 pages

Solution 4

The document discusses multivariate Gaussian distributions and their properties. It covers topics like the covariance matrix of a random vector, orthonormal basis vectors of a subspace, the variance of projecting a vector onto another vector, how properties change when a linear transformation is applied to a Gaussian, and formulating a classifier to choose between two Gaussian distributed classes.

Uploaded by

chandraadityaoffl
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
31 views2 pages

Solution 4

The document discusses multivariate Gaussian distributions and their properties. It covers topics like the covariance matrix of a random vector, orthonormal basis vectors of a subspace, the variance of projecting a vector onto another vector, how properties change when a linear transformation is applied to a Gaussian, and formulating a classifier to choose between two Gaussian distributed classes.

Uploaded by

chandraadityaoffl
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

1.

Consider a random column vector 𝐱̅ that has the multi-variate Gaussian distribution
with mean 𝛍 ̅. Its covariance matrix is defined as
̅ )𝑇 }
̅)(𝐱̅ − 𝛍
𝐸{(𝐱̅ − 𝛍
Ans c
2. The picture shown corresponds to a Classifier
Ans b
3. A set of orthonormal basis vectors for the same subspace is given as
1 −3
1 1 1 −1
𝐯̅1 = [ ] , 𝐯̅2 = [ ]
2 1 2√5 1
1 3
Ans d
4. The variance of the quantity 𝐚̅𝑇 𝐱̅ is
𝜎 2 ‖𝐚̅‖2
Ans c
5. The vector 𝐀𝐱̅ + 𝐛̅ is Gaussian with mean and covariance
𝐀𝝁̅ + 𝐛̅, 𝐀𝚺𝐀𝑇
Ans d
6. The Gaussian classification problem with the two classes 𝐶1 , 𝐶2 distributed as
1 1
0 0
−1 2
𝐶1 ∼ 𝑁 ([ ] , [2 ]) , 𝐶2 ∼ 𝑁 ([ ] , [2 ])
2 1 1 1
0 0
4 4
The classifier chooses 𝐶1 if
𝛍
̅1 + 𝛍̅2
(𝛍
̅1 − 𝛍̅2 )𝑇 𝚺 −1 (𝐱̅ − )≥0
2
−3 𝑇 2 0 1 1
⇒ ([ ]) [ ] (𝐱̅ − [ ]) ≥ 𝟎
1 0 4 2 3
1 1
⇒ [−6 4] (𝐱̅ − [ ]) ≥ 𝟎
2 3
⇒ −6𝑥1 + 4𝑥2 ≥ 3 ⇒ 6𝑥1 − 4𝑥2 ≤ −3
Ans a
7. The probability distribution function (PDF) 𝑓𝑋 (𝑥) is given as
1 (𝑥−𝜇)2
−
𝑒 2𝜎2
√2𝜋𝜎 2
Ans c
8. The classifier for this problem, to classify a new vector 𝐱̅, can be formulated as
𝛍
̅1 +𝛍
̅2
Choose 𝐶1 if (𝛍 ̅2 )𝑇 𝚺 −1 (𝐱̅ −
̅1 − 𝛍 ) ≥ 0 and 0 otherwise
2
Ans b
9. The eigenvectors of the matrix are
3 1
[ ],[ ]
2 −1
Ans c
10. Given
1 −2
𝐱̅1 = [1] , 𝐱̅ 2 = [ 1 ]
1 2
1 −3
The orthonormal vector basis can be found as follows
1
𝐱̅1 1 1
𝐯̅1 = = [ ]
‖𝐱̅1 ‖ 2 1
1
−2 1
1
𝐯̃2 = 𝐱̅ 2 − 𝐯̅1 𝐯̅1𝑇 𝐱̅ 2 = [ 1 ] − [1] × (−2)
2 4 1
−3 1
3
−
2
−2 1 3 −3 −3
1 1 1
= [ 1 ] + [1] = 2 = [ 3 ] = [3]
2 2 1 5 2 5 2√17 5
−3 1 2 −5 −5
5
−
[ 2]
Ans d

You might also like