Basic question

LevyProcess · February 2nd, 2005, 4:48 pm

Hi everyone, i was wondering if anyone can assist me solve the following problem.I have one dependent variable Y and 5 independent variables X1, X2, X3, X4, X5.I want to determine the contribution of each independent variable to the realisation of Y. I believe that a multinomial logit model should enable me to do this. Can anyone explain how this works and how to implement it in SAS or SPSS?Thank you.

exotiq · February 2nd, 2005, 5:13 pm

What are you doing that's different than a straight 5-factor linear regression you can do with linest() in Excel?

SPAAGG · February 2nd, 2005, 6:22 pm

I agree with exoticCan you give use the type of variable you have (i.e dummy, continus, etc...). It might help us to give you an answer.You can also take a look at Principal Component Analysis

LevyProcess · February 2nd, 2005, 7:43 pm

Ok exotiq and spaagg,assume that I want to look at how the following variables: Fico score, loan-to -ratio, coupon type ( fixed or ARM), loan balance, and geographical location impact the probability that a borrower would default on a loan. What would be the best model to perform that task?Thanks.

SPAAGG · February 3rd, 2005, 9:06 am

In that case, your multinomial logit model seem appropriate.However, I don't know the theory behind it. Can't help you more...

silvione · February 3rd, 2005, 10:01 am

try these sites:http://www.washington.edu/computing/sof ... ds13.0.pdf or http://www.indiana.edu/~statmath/stat/all/cat/2c1.html or http://www.urbansim.org/class/presentat ... cture8.pdf or http://www.siue.edu/IUR/SPSS/SPSS%20Reg ... 2012.0.pdf.

GregWallace · February 24th, 2005, 1:06 pm

Please provide some message in your topic heading so we know what you are asking about prior to digging deeper.Not doing this will lose most readers.