fitnlm and dummy variables

2 views (last 30 days)
Tania
Tania on 18 Jun 2014
Commented: the cyclist on 23 Jun 2014
Hi all, Fitnlm allows automatic creation of dummy variables, if one input the categorical predictor as a nominal or ordinal array. It worked fine for me with fitlm. But I do not understand how I know how many input arguments I get after using the automatic creation of dummy variables. As I need to specify my beta0 I need to know how many input arguments I have. Thank you!

Accepted Answer

the cyclist
the cyclist on 18 Jun 2014
Edited: the cyclist on 18 Jun 2014
I haven't used the automatic creation of dummy variables, so I don't know the answer to your question. However, the dummyvar() function might also be helpful for you, so I thought I'd mention it.
  3 Comments
Tania
Tania on 23 Jun 2014
Hi, I have created a dummyvar for my postcode variable now: dv=dummyvar(B) But I am not sure how I use this now in fitlm (or fitnlm). Without dummyvar I would do the following: >> ds = dataset(price, bedroom, school, transport); >> mdl =fitlm(ds, 'price~bedroom + school + transport') How do I include my dv now into my dataset? My dv has 29 columns. Thank you! PS: I have my document attached!
the cyclist
the cyclist on 23 Jun 2014
Suppose you have the following grouping variable (where "1", "2", etc represent the groups):
group = [1;1;1;2;2;3;3;3;3;4]
Then
dv = dummyvar(group)
gives
dv =
1 0 0 0
1 0 0 0
1 0 0 0
0 1 0 0
0 1 0 0
0 0 1 0
0 0 1 0
0 0 1 0
0 0 1 0
0 0 0 1
Each column of dv is a variable in your regression. The first column is the binary variable "is_member_of_group?", where 1=yes and 0=no. Likewise for other columns.

Sign in to comment.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!