Loading variables & using 'for' loop for comman polyeig()

Let us say, i've a series of matrices as K1, K2,....K(n)

I want to obtain eigen values and eigen vectors for the above mentioned matrices; like:

[X1 E1] = polyeig(K1,M)

[X2 E2] = polyeig(K2,M)

[X3 E3] = polyeig(K3,M) & so on

Can anyone suggest a for loop for it, where M is not a variable matrix. But X1,X2,X3...... & E1,E2,E3..... are variables.

Answer 1

[Ameer Hamza]

It is not a good idea to name variable like K1, K2, K3, ... It makes code complicated: https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval. First I will convert K1, K2, ..., Kn to cell array and then return the output in cell arrays.

K = {K1, K2, K3, .., K10}; % list all the Kxx variables here
M; % matrix
[X, E] = cellfun(@(k) polyeig(k, M), K, 'uni', 0);

Access the results like this

X{1}; % first element of X
X{2}; % second element of X
..
X{end}; % last element of X
E{1}; % first element of E
E{2}; % second element of E
..
E{end}; % last element of E

[Parvesh Deepan]

Hi Ameer,

I really want to thank you for giving your valuable time on it and giving this solution. It is showing some error for me, i don't know why. So, I'm sharing mine code, so that you will be in better position to answer the problem.

close all

clc

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

m1 = 16500;

m2 = 16100;

M = [m1 0;0 m2];

x1 = linspace(0,3,5000);

x2 = linspace(0,2,5000);

k1 = 2.968888888888889e+07; %(k1=k2)

Ks1 = k1*x1;

Ks2 = k1*x2;

for i = 1:length(x1)

eval(['K' num2str(i) '= [Ks1(1,i)+Ks2(1,i) -Ks2(1,i);-Ks2(1,i) Ks2(1,i)];']);

end

%% I need to evaluate:

% [X1,e1] = polyeig(K1,M)

% [X2,e2] = polyeig(K2,M) & so on...

% where X represents Eigen-Vectors & e represents Eigen-values

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

[Ameer Hamza]

Run the following code. It will be mush faster as compared to the current code (Remember, eval is evil).

close all
clc
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
m1 = 16500;
m2 = 16100;
M = [m1 0;0 m2];
x1 = linspace(0,3,5000);
x2 = linspace(0,2,5000);
k1 = 2.968888888888889e+07; %(k1=k2)
Ks1 = k1*x1;
Ks2 = k1*x2;
K = cell(size(x1));
for i = 1:length(x1)
   K{i} = [Ks1(1,i)+Ks2(1,i) -Ks2(1,i);-Ks2(1,i) Ks2(1,i)];
end
[X, E] = cellfun(@(k) polyeig(k, M), K, 'uni', 0);

[Parvesh Deepan]

Thanks a lot @Ameer Hamza. It really works.

And that was amazing (eval is evil) :)

[Ameer Hamza]

I am glad to be of help!

[Parvesh Deepan]

I've a MATLAB code to obtain prior PDF plot, as:

%%------------------------------------------------------------------------------------------------

% Prior PDF--

theta1 = linspace(0,3,5000);

theta2 = linspace(0,2,5000);

p0 = lognpdf(theta1,1.3,1).*lognpdf(theta2,0.8,1); % Prior-PDF function

syms t

g1 = (1/t)*exp(-0.5*((log(t)-1.3)^2));

g2 = (1/t)*exp(-0.5*((log(t)-0.8)^2));

a = int(g1*g2,t,0,3);

b = (1/(2*pi()))*double(a);

c = 1/b;

y = c*p0;

figure

plot(theta1,y,'k','LineWidth',2)

grid on;

axis([0 3 0 1])

xlabel('Theta-1');

ylabel('Prior PDF');

title('Prior-PDF function Plot-1');

%%-------------------------------------------------------------------------------------------%%

Now, i want the same results using ksdensity function? How to obtain this?

Can you answer this?