Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "ele.mose@libero.it" <ele.mose@libero.it> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: info stata |
Date | Fri, 27 May 2011 12:47:08 +0200 (CEST) |
Dear sirs, I'm a master degree stundent and I'm analysing a dataset of a credit agency, for the aim of developing a credit scoring model (using Stata for logistic regression). I'm having trouble for two points and I will be grateful if you could help me: 1. I have splitted my dataset in two samples: validation and training. After having developed my model on the training and tested the performance trough the estat classificacion, how do I work on the validation one? should I launch on it my model (logit launched on the training sample) and test its performance trough the estat, then if it is ok continuine to work on the training sample, to establish predicting probabilities of defaut (that is my dependent var: 0 or 1)? or should I work on the validation sample, after having tested its performance? 2. does it exist a command in stata that enable me to list, other than the classification table, both sensitivity and specificity for each observation in the dataset? Thank you very much for your attention, Eleonora Moselli. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/