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];
beta = regress(X,Y);
but it would return the exact same thing as
beta = X\Y;
