Please help me with my code ( linear regression)

11 次查看（过去 30 天）

Mark Lu 2017-3-30

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/332863-please-help-me-with-my-code-linear-regression

回答： Nicolas Schmit 2017-9-5

    if true
      % code
    endclear; clc;
  %Step1:Generate a vector Y and X
  n=1000;
  v=2;
  Y=trnd(v,n,1);
  x1=ones(n,1);
  x2=randn(n,1);
  X=[x1 x2]; 
%Step2:Compute the OLS estimator
beta=(X'*X)^(-1)*X'*Y;
%Step3: Repeat 1. and 2. MC = 10000 times.
MC= 10000;
for nsample = 1:5
  if nsample == 1; n = 1000;  end;
  if nsample == 2; n = 2000;  end;
  if nsample == 3; n = 3000;  end;
  if nsample == 4; n = 5000;  end;
  if nsample == 5; n = 7000;  end;
    for n=1:MC;
      beta_hat(n,:) = (X'*X)^(-1)*X'*Y
      for df=1:10;
          B0=beta_hat(n,:)- beta;
          bias=(1/MC)*symsum(B0,n,1,MC);
          rmse=sqrt(bias^2);
      df=[1:0; 1.1; 1.2; 1.3; 1.4; 1.6; 1.8; 2.0; 5.0; 10.0];
  end;
  end;
  end;

The statements I am trying to code is below

5 个评论
显示 3更早的评论隐藏 3更早的评论

Mark Lu 2017-3-31

Here is the full error message . ''Error using - Matrix dimensions must agree. Error in Assign3 (line 27) B0=beta_hat(n,:)- beta;''

Image Analyst 2017-4-3

What are the sizes of beta_hat and beta? How many rows and columns? Evidently they're not the same! beta_hat(n, :) is a 1-D row vector and you're probably trying to stuff a 2-D matrix into a single row and it won't fit.

请先登录，再进行评论。

请先登录，再回答此问题。

回答（1 个）

Nicolas Schmit 2017-9-5

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/332863-please-help-me-with-my-code-linear-regression#answer_280304

The error message means that you are trying to subtract two variables with incompatible sizes. This occurs because size(beta)= [2 1] whereas size(beta_hat(n,:)) = [1 2]. Transpose beta to make it a row vector.