Forum: help

RE: z variables in sfa function [ Reply ] By: Arne Henningsen on 2017-06-10 05:13	[forum:45193]
You can create a 'minimal reproducible example' by reducing the number of explanatory variables and the number of observations. Creating a 'minimal reproducible example' often helps to find the problem, because you often figure out, which variables and/or observations create the problem.

RE: z variables in sfa function [ Reply ] By: Yvette Zen on 2017-06-01 08:10	[forum:45173]
honestly saying, i cant imagine how to do it. because i have big dataset. but anyway i solved all my problems. thanks again for your responsiveness!!!

RE: z variables in sfa function [ Reply ] By: Arne Henningsen on 2017-05-28 15:51	[forum:45164]
Okay. If you post a minimal example that reproduces your error message, I can fix the error -- or at least let sfa() return a more helpful error message. /Arne

RE: z variables in sfa function [ Reply ] By: Yvette Zen on 2017-05-26 12:53	[forum:45162]
thank you for your answer! the thing is exactly if i log one of them (z variable that has big values) or all, then everything is good and no errors! best regards

RE: z variables in sfa function [ Reply ] By: Arne Henningsen on 2017-05-25 05:37	[forum:45159]
You can use logged variables as z variables as long as the original variables are strictly positive (given that the logarithms of zeros and of negative values are undefined). I am not sure why you get the error message if you use logged z variables. Can you please provide a minimal reproducible example?

z variables in sfa function [ Reply ]
By: Yvette Zen on 2017-05-17 16:29

[forum:45134]

Dear Mr. Henningsen,

i am sorry for being intrusive here, but one more question, please

i dont understand from the mathematical point (unfortunately) if it is possible to log z variables in sfa function ?
because if i dont log one of z variables (inefficiency), R says me the next:
"function value in search procedure is NaN
Error in if ((returnObj$mleParam["gamma"] < 0.01 || returnObj$mleParam["gamma"] > :
missing value where TRUE/FALSE needed"

thank you !