Re: вот вроде как инструкция по заскриненым комментарияp2004rMarch 2 2015, 13:02:49 UTC
Я то объясню...
> choose.direction(rsmax2(data1, optimized=F), data = data1, c("obe1", "black"), criterion = "bic-g", debug = TRUE) * testing obe1 - black for direction. > initial score for node obe1 is -122.4541 . > score delta for arc obe1 -> black is 10.61515 . > initial score for node black is -189.4856 . > score delta for arc black -> obe1 is 1.06267 . @ arc obe1 -> black is better . > choose.direction(rsmax2(data1, optimized=F), data = data1, c("obe1", "black"), criterion = "aic-g", debug = TRUE) * testing obe1 - black for direction. > initial score for node obe1 is -120.5223 . > score delta for arc obe1 -> black is 11.58106 . > initial score for node black is -186.5878 . > score delta for arc black -> obe1 is 2.028583 . @ arc obe1 -> black is better . > choose.direction(rsmax2(data1, optimized=F), data = data1, c("obe1", "black"), criterion = "bge", debug = TRUE) * testing obe1 - black for direction. > initial score for node obe1 is -122.5446 . > score delta for arc obe1 -> black is 8.836843 . > initial score for node black is -189.0292 . > score delta for arc black -> obe1 is 1.507393 . @ arc obe1 -> black is better . >
Re: вот вроде как инструкция по заскриненым комментарияsecondary_teaMarch 17 2015, 09:09:51 UTC
Я поставил bnlearn себе, скажите, как воспроизвести ваши результаты Ну вот вы привели вывод внутренней кухни, это конечно интересно, но не настолько, как выводы этого алгоритма -- как интерпретировать obe1 -> black?
> choose.direction(rsmax2(data1, optimized=F), data = data1, c("obe1", "black"), criterion = "bic-g", debug = TRUE)
* testing obe1 - black for direction.
> initial score for node obe1 is -122.4541 .
> score delta for arc obe1 -> black is 10.61515 .
> initial score for node black is -189.4856 .
> score delta for arc black -> obe1 is 1.06267 .
@ arc obe1 -> black is better .
> choose.direction(rsmax2(data1, optimized=F), data = data1, c("obe1", "black"), criterion = "aic-g", debug = TRUE)
* testing obe1 - black for direction.
> initial score for node obe1 is -120.5223 .
> score delta for arc obe1 -> black is 11.58106 .
> initial score for node black is -186.5878 .
> score delta for arc black -> obe1 is 2.028583 .
@ arc obe1 -> black is better .
> choose.direction(rsmax2(data1, optimized=F), data = data1, c("obe1", "black"), criterion = "bge", debug = TRUE)
* testing obe1 - black for direction.
> initial score for node obe1 is -122.5446 .
> score delta for arc obe1 -> black is 8.836843 .
> initial score for node black is -189.0292 .
> score delta for arc black -> obe1 is 1.507393 .
@ arc obe1 -> black is better .
>
...ну вот, объяснил, стало легче? :)
Reply
Ну вот вы привели вывод внутренней кухни, это конечно интересно, но не настолько, как выводы этого алгоритма -- как интерпретировать obe1 -> black?
Reply
Reply
Reply
Reply
Leave a comment