bug in multinomial GMPMs
Monitor Forum | 
Start New Thread|
bug in multinomial GMPMs By: Dale Barr on 2009-08-31 08:45 | [forum:1868] |
When you are fitting a multinomial GMPM, and press the relevant key command to stop the fitting (e.g., Ctrl-C), R seems to become unresponsive to user input. WORKAROUND: a call to "sink(NULL)" (or two) after breaking restores the output to the command terminal The reason for this behavior is that within the fitting subroutine, which calls the R function "multinom" from package nnet, the output of nnet is redirected to a file called ".multinom.txt" using the sink() function. This was to avoid cluttering up the display. In each run, before the call to multinom(), a call is made to the sink() function. If the user breaks while the multinom function is running, then the "unsink" command never completes, and so output continues to be re-directed to the file. | |
Thread View
| Thread | Author | Date |
|---|---|---|
 bug in multinomial GMPMs | Dale Barr | 2009-08-31 08:45 |
