linear regression with matrix

Hi guys, I am new to mablat. I have 4 4424x2380 matrixs and I want to analyst every single point of the matrix. That means I need to have a equation for, for instance, colume 3 row 3 number for 4 matrix for further use. Can anyone know how to do it? No worry about the computing power.

[Ho Leung Louise FUNG]

for instance

a= [2 2 5]

[13 2 6]

[3 3 9]

b= [5 6 1]

[2 14 53]

[6 12 42]

I need a linear equation for a11 and b11 and a22 and b22

I guess it is kind of matrix transformation and fitlm() may help me.

But fitlm require one of the matrix be single

[Torsten]

I need a linear equation for a11 and b11 and a22 and b22

Unclear what you mean.

[Star Strider]

What is being regressed with respect to the independent and dependent variables?

[John D'Errico]

A linear regression requires an independent variable, AND a dependent variable. It seeems your dependent variable may be the numbers contained in these 4 matrices. PLEASE LEARN TO USE MATRICES PROPERLY. @Atsushi Ueno has suggested you should be storing these numbers in one array, of size 4424x2380x4. That is a wonderful idea. But it still does not answer the relevant question, that is, what is the INDEPENDENT variable in the regression you want to perform? Essentially, it looks like y is a vector of length 4. But what is x?

Answer 1

[Atsushi Ueno]

How about stacking them as a 4424x2380x4 matrix?

A = ones(4424,2380);
B = ones(4424,2380);
C = ones(4424,2380);
D = ones(4424,2380);
E = cat(3,A,B,C,D);
size(E)

ans = 1×3

4424 2380 4

E(3,3,:) % for instance, colume 3 row 3 number for 4 matrix for further use.

ans = 
ans(:,:,1) =

     1


ans(:,:,2) =

     1


ans(:,:,3) =

     1


ans(:,:,4) =

     1