12-14-2011 11:09 AM
I am running -mprobit- on -proc mdc-. It gives me the note: At least one element of the (projected) gradient is greater than 1e-3. And, results of some variabels vanishes. Can anybody help me how to resilve this problem?
12-15-2011 10:36 AM
Hi everybody, I had posted a question yesterday. It was not responded aby anyone as I think it was incomplete. In addition to the above question, there are more queries.
As I get the result from "Proc mdc" for a multinomial probit model for most of the parameter estimates the DF is 0 "instead of 1" and there is no result for standard error and t value. And the note: At least one element of the (projected) gradient is greater than 1e-3.
can anybody help out?