I used a multinomial logistic regression on a csv file (3400 records) and the continuous independent variables has too many different variables (2500).
When I run the multinominal regression model it takes too much time and then failed to result.
I also assigned the MaxNWts = 1000000.
Do you have any idea?
Asked
Active
Viewed 29 times
0
-
As you pointed out, you have too many categories for the model to estimate. – Phil Nov 17 '21 at 20:06
-
is there any solution for not getting failed? – Emma Nov 17 '21 at 21:04
-
Reduce the number of categories by grouping them up in some way, or get more data with enough observations in each of the categories. – Phil Nov 17 '21 at 21:59