multiple linear regression check

Hello Senior Matlab users

Firstly, I want to say sorry for my question as I am a beginner.

I want to ask for good programming advice for my messy code.

In my multiple linear regression example, there are two dependent variables( wind, temperature) and independent variable(evporation) for 10 numbers.

I got coefficients of beta1, beta2 and beta3 and I also checked the answer using calculated beta values. The answer seems correct.

The question is I used many for loops and I don't know how to code in only one for loop and smarter way. I also want to use built in regression function but don't know how to include.

So, if you can advice for my messy code, it would be very helpful for me.

Thank you very much.

clear all;clc;
data = csvread('multiregression.csv');
X = data(:,2:4);
wind= X(:,2);
temp= X(:,3);
Y = data(:,1);
Xt = data(:,2:4)';
Yt = data(:,1)';
beta= ((Xt*X)^-1)*(Xt*Y)

X2= size(X,1);
for i=1:X2
    Xq(i)=beta(2)* wind(i);
end

for i=1:X2
    Yq(i)= beta(3)*temp(i);
end
beta1= (beta(1)* ones(X2,1));
for i=1:X2
    calculated_ans(i)= beta1(i) + Xq(i) + Yq(i);
end
answer= [calculated_ans' Y]

Answer 1

[Tommy]

All of the loops can be avoided. If you have something like

b=1:10;
for i=1:numel(b)
    a(i)=2*b(i)+3;
end

you can always just use

b=1:10;
a=2*b+3;

instead, and MATLAB will apply the operations to each element within b. The following gives the same result as your code:

data = csvread('multiregression.csv');
Y = data(:,1);
X = data(:,2:4);
wind = X(:,2);
temp = X(:,3);
beta = X\Y;
calculated_ans = beta(1) + beta(2)*wind + beta(3)*temp;
answer = [calculated_ans Y];

You could use the MATLAB function regress, like so

beta = regress(X,Y);

but it would return the exact same thing as

beta = X\Y;

[soe thiha]

Hello Tommy

Thank you very much for your suggestion and it is really helpful for me.

Regards,