Could anyone explains what is the error and how could I resolve it?
> C.m<-mlogit(choice~TotalAmount,data=C)
Error in solve.default(H, g[!fixed]) :
system is computationally singular: reciprocal condition number = 6.87101e-23
Asked
Active
Viewed 1,277 times
2
smci
- 32,567
- 20
- 113
- 146
chee.work.stuff
- 326
- 2
- 14
-
Added better title and tagged [tag:mlogit]. Also, tell us your package versions, upgrade to the latest, open a clean R session, reverify you still see your issue. Mine are mlogit 0.2.4 (maxLik 1.2.0). – smci May 18 '14 at 18:46
-
And it goes without saying: **give us a reproducible code snippet, use `dput` (don't dump everything!) and `str(C.m)`** so we understand what type each of the columns of *`C.m`* are - your question currently says nothing at all about that. – smci May 18 '14 at 18:47
1 Answers
3
Something in your data is computationally singular e.g. some of the columns you are trying to do the regression on are correlated.
Maybe you can remove some of the correlated columns and try again without them.
doctorlove
- 18,872
- 2
- 46
- 62
-
I have filtered the variables with correlation > 0.91, but the error message remain. – pengchy Jun 28 '16 at 04:35