[Limdep Nlogit List] Nlogit crosstabs Slight change in what follows.
William Greene
wgreene at stern.nyu.edu
Tue Aug 24 14:22:04 EST 2010
Cathy.
The P(j,t) is the fitted probability for outcome j for person t.
The Y(j,t) is the dummy variable that indicates if individual
t chose alternative j or not. So, Y(j,t) equals 0 for J-1 of the
j's and 1 for the j that corresponds to the choice made.
/B. Greene
----- Original Message -----
From: "Cathy Durham" <cathy.durham at oregonstate.edu>
To: "Limdep and Nlogit Mailing List" <limdep at limdep.itls.usyd.edu.au>
Sent: Monday, August 23, 2010 7:49:33 PM GMT -05:00 Colombia
Subject: [Limdep Nlogit List] Nlogit crosstabs Slight change in what follows.
I've come to the realization that the actual v. predicted table produced from the ;crosstab statement with an nlogit run is not what I expected. I reproduced the values resulting from
;list
by writing out the formulas to create predictions and they are what I expected, but using the same principles to replicate the crosstab output does not. One would think the correctly predicted numbers would be the same.
The Nlogit 3.0 reference manual (p.N3-21) says the predictions are obtained as the Integer part of the formula Sum over t P^ sub jt times y sub jt
What exactly are P^ and y in this formula?
Is t the choice set, and is j the alternative?
In the printout tabulations are headed with the following.
| Cross tabulation of actual vs. predicted choices. |
| Row indicator is actual, column is predicted. |
| Predicted total is F(k,j,i)=Sum(i=1,...,N) P(k,j,i). |
| Column totals may be subject to rounding error. |
Thanks.
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au
More information about the Limdep
mailing list