计算平方残差之和？

Question

我正在尝试编写一个代码，用于相对于糖和纤维的营养评级进行回归。并计算平方残差之和R2和估计的标准误s。

我是Matlab的新手，但一直遵循https://www.mathworks.com/help/matlab/data_analysis/linear-regression.html

中的方法

  z=csvread('Cereals no alpha.csv');
    [rows,cols]=size(z);
    disp([rows,cols])
    sugar=z(:,7);
    fiber=z(:,5);
    rating=z(:,13);

    %Regression
    V=ones(rows,3);
    V(:,2)=sugar;
    V(:,3)=fiber;
    A=V'*V;
    b=V'*rating;
    w =A\b;

    b0=w(1);
    bs=w(2);
    bf=w(3);
    disp([b0,bs,bf])

    %Sum squared error performance function
    %perf = sse(z,sugar,fiber);
    %disp(perf)


    %R^2 maybe?
    %Use polyfit to compute a linear regression that predicts y from x:
    p = polyfit(sugar,fiber,1)

    %fit equation
    yfit =p(1)*sugar+p(2);

    %Compute the residual values as a vector of signed numbers:
    yresid = fiber-yfit;

    %Square the residuals and total them to obtain the residual sum of 
    SSresid =sum(yresid.^2);

    %Compute the total sum of squares of y by multiplying the variance of
    by the number of observations minus 1:
    SStotal = (length(fiber)-1) * var(fiber);

    %Compute R2
    rsq = 1 - SSresid/SStotal

给予： disp（[b0，bs，bf]）= 51.7635 -2.2012 2.8661

p = -0.0761 2.6797

rsq = 0.0199

但是根据数据，我预计回归的R2约为80.8％，s = 6.24