you can pretend
p(e|f) = exp(-465.62)
this is what other algorithms such as Minimum Bayes Risk do.
On 05/05/2014 05:30, arpit gupta wrote:
0 ||| move blockInRoom shapeChair agentInRoom bump blockInRoom
roomIsGreen blockIsYellow colored area ||| d: -12 -0.510826 -1.60944
-1.35635 -0.685179 -0.251314 -1.53248 lm: -95.8405 -145.991 w: -10 tm:
-27.0221 -18.3061 -4.01228 -5.55009 5.99938 ||| -465.62
I am trying to get probability like p(e|f) of sentences . consider the
translated sentence to be in english language , in this sample. I am
not able to see how to transform these score into actual probability
or probability like scores.
Thanks
Arpit
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
