trying to do anova1 for multiple classes

i have 4 classes of size (72,22,22) and i want to do anova1 for each pair of the (22,22) in those classes .

for example; i wanna grab, let's say, pair (2,3) in each class and do the anova for each pair in the 4 arrays, so the output would be an array of (22,22)representing P-values of each pair across 4 classes. hopefullay my code explains what i am trying to say :)

i have tried this piece of code but im not sure cuz the results are kinda odd; i guess the issue is in 2 for loops. here is the code

>> x = load("plv_8_12.mat");
x = x.plv;
size(x)
ans =
        2592          22          22
S1C1 =x(1:72,:,:);
S1C2 = x(649:720,:,:);
S1C3 = x(1297:1368,:,:);
S1C4 = x(1945:2016,:,:);

p_all = [];
for x =1:22
    for y=1:22
        
        tc1 = S1C1(:,x,y);
        tc2 = S1C2(:,x,y);
        tc3 = S1C3(:,x,y);
        tc4 = S1C4(:,x,y);
        temp = [tc1 tc2 tc3 tc4];
        
        %p_all(x,y) = anova1(temp);
        [p,tbl,stats]=anova1(temp);
        
        close all
    end
end

Answer 1

[Shivansh]

Hi Abdullah,

I understand that you want to get results with anova1 function for each pair of the classes defined by you.

Your code looks fine, but there is a small issue with the way you are storing the p-values. I believe that the issue lies in storage of p-values for each pair of (x,y) in the loop. The 'p' variable is getting overwritten in each iteration of the loop instead of being stored. To store the p-values for each pair, you can modify your code as follows:

x = load("plv_8_12.mat"); 
x = x.plv; 
size(x) 
ans = 
2592          22          22 

S1C1 = x(1:72,:,:); 
S1C2 = x(649:720,:,:); 
S1C3 = x(1297:1368,:,:); 
S1C4 = x(1945:2016,:,:); 

p_all = zeros(22, 22);  % Initialize p-values matrix 

for x = 1:22 
    for y = 1:22 
        tc1 = S1C1(:,x,y); 
        tc2 = S1C2(:,x,y); 
        tc3 = S1C3(:,x,y); 
        tc4 = S1C4(:,x,y); 
        temp = [tc1 tc2 tc3 tc4]; 
        [p, ~, ~] = anova1(temp); 
        p_all(x, y) = p;  % Store p-value in the matrix 
    end 
end 

In this modified code, I added a `p_all` matrix to store the p-values for each pair of (x, y) in the loop. The `p_all` matrix is initialized with zeros, and in each iteration of the loop, the corresponding p-value is stored in the matrix using the indices (x, y). This way, you will have the p-values for each pair of (x, y) across the four classes in the `p_all` matrix.

I hope it resolves your issue!