# Thread Subject: Binary logit regression

 Subject: Binary logit regression From: Victor Date: 30 Jun, 2012 17:10:15 Message: 1 of 3 Good afternoon, I am attempting to model sovereign default probabilities. To do so I have as dependant variable: y=[0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1] and four explanatory variables x1 x2 x3 and x4, all of same dimension. Here are my issues: 1) Shall I regress first my explanatory variables against the variable Y (which is not in a binary form) from which I derived y? Or should I directly regress these explanatory variables against y? If the second option is the right one, how do I do that? 2) I do not really know what are the steps I should follow then to obtain the probability of default, according to the variable that I will find significant. Assuming x1 and x2 are significant, are the following steps correct? X=[x1 x2] [b,dev,stats] = glmfit(X,y,'binomial','link','logit') yfit = glmval(b, X,'logit'); 3) Finally, if the above steps are correct, among the outputs returned, which one would give me the probability of default? Thanking you in advance for your help. Victor
 Subject: Binary logit regression From: Greg Heath Date: 1 Jul, 2012 05:25:24 Message: 2 of 3 On Jun 30, 1:10 pm, "Victor " wrote: > Good afternoon, > > I am attempting to model sovereign default probabilities. To do so I have as dependant variable: > y=[0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1] > and four explanatory variables x1 x2 x3 and x4, all of same dimension. > > Here are my issues: > 1) Shall I regress first my explanatory variables against the variable Y (which is not in a binary form) from which I derived y? Or should I directly regress these explanatory variables against y? > > If the second option is the right one, how do I do that? > > 2) I do not really know what are the steps I should follow then to obtain the probability of default, according to the variable that I will find significant. > > Assuming x1 and x2 are significant, are the following steps correct? > > X=[x1 x2] > [b,dev,stats] = glmfit(X,y,'binomial','link','logit') > yfit = glmval(b, X,'logit'); > > 3) Finally, if the above steps are correct, among the outputs returned, which one would give me the probability of default? yfit. help glmfit doc glmfit Hope this helps. Greg
 Subject: Binary logit regression From: Victor Date: 1 Jul, 2012 10:04:07 Message: 3 of 3 Hi Greg, Thanks for your prompt answer. I unfortunately already read the matlab doc as well as other posts in matlab central but I am still not sure of what I have to do. If you could give me some advices regarding the above mentioned three steps, I would really appreciate. Thanks again for your help. Victor

## Tags for this Thread

### Add a New Tag:

Separated by commas
Ex.: root locus, bode

### What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Tag Activity for This Thread
Tag Applied By Date/Time
logit Victor 30 Jun, 2012 13:14:13
binary Victor 30 Jun, 2012 13:14:13
Feed for this Thread