ASSume WT hawt Fhree classes, aneh Huo Adve nsionnk Baitinghoot Ah Follow ney means and Covacianet...

Question

ASSume WT hawt Fhree classes, aneh Huo Adve nsionnk Baiting
hoot Ah Follow ney means and Covacianet makaies :
A mela, My = I! ; My = 15]
p h I~)
2-2 ne [0] ;
poy Probabiliticy, J? (»)) = p (<2) z= Are J I? (Ws) z =
0 } fname Als lrd mand fanctim¢ 0),(%),9 &) , (x) , Hen
Paph +e decision bowndacy rig (ens -
2- ¢ loss fy a Fedurt Veet rely
Wy z wy) = PR = L
& 2 pled Aha deciscan bowndary ASSUmIA pon =y! D ¥Y 3) 2
- re peat pact o a SSA Mey Fl covariant mates a~
Ack ferent : ~*% A i)
3
\ % rh 2 F / \ 5
2 z I} i.
Find 90 a0, 900 asuming p(a)=P@) = 5, F233
Jas Ha gecime besedee) s BU 5002 9,00
ALr = A,
e lass fo a de don yor Re (5) .
9

screenshot-2022-10-26-at-95754-am-0h0qq5wd.png

Rhea · Accepted Answer

Given, we have three classes, each 2d Gaussian having the following means and covariance matrices-
 and 
Prior probabilities: 
Let, 
We need to-
1. Find linear discriminant functions  and then graph the decision boundary regions.
The linear discriminant function for each class is-
Since, , so the covariance is class-independent and hence we can drop the terms which do not contain mean or prior probabilities as irrelevant.
Similarly, we can find  and . Solving for separately for  and  we’ll get the decision boundary.
Using Matlab, we obtain the following-
mu1=[1;3];
mu2=[-1;6];
mu3=[-2;5];
sigma1=[1,-1;-1,4];
sigma2=[1,-1;-1,4];
sigma3=[1,-1;-1,4];
p1=1/5;
p2=1/5;
p3=3/5;
syms x1 x2;
g1 = -0.5*([x1;x2]-mu1)'*[inv(sigma1)*([x1;x2]-mu1)]-log(2*pi)-0.

ASSume WT hawt Fhree classes, aneh Huo Adve nsionnk Baiting hoot Ah Follow ney means and Covacianet makaies : A mela, My = I! ; My = 15] p h I~) 2-2 ne [0] ; poy Probabiliticy, J? (»)) = p (

Solution

Answer To This Question Is Available To Download

Related Questions & Answers

Submit New Assignment