MATLAB Answers

Hinti123
0

Regression using 1x6 table with 500x20 double elements

Asked by Hinti123
on 9 Nov 2018 at 20:49
Latest activity Commented on by Hinti123
on 11 Nov 2018 at 23:30
Hi all, i want to do a linear regression using fitlm. I created a table with 6 variables. The table consints of one row where each variable stores 500x20 double elements. I now want to do a linear regression with Var1 as Y and Var2-Var6 as X. So it should look like this:
Var1(row,i)=(alpha)+(beta1)*Var2(row,i)+(beta2)*Var3(row,i)+...
This has to be done for each row separate. So in the endeffect I should have 1 alpha and 5 Betas for 500 rows. That is my regression command i am trying to execute
fitlm(Mytable,'Var1~Var2+Var3+Var4+Var5+Var6');
But i keep getting this error
Error using classreg.regr.FitObject/selectVariables (line 293)
Predictor variables must be numeric vectors, numeric matrices, or categorical vectors.
Can you please help me? Thanks

  0 Comments

Sign in to comment.

1 Answer

Answer by Jeff Miller on 11 Nov 2018 at 0:01
 Accepted Answer

I'm not sure I fully understand your question, but maybe this will help: If you want one alpha and five betas for each of the 500 rows separately, then you will have to call fitlm 500 times. Each time you call it, you should pass only the data for one row.

  1 Comment

ok. Yeah, that is exactly what I was expecting. Thanks for your help

Sign in to comment.