Stochastic context-free grammars to Chomsky normal form

Question

I actually don't understand how to convert scfg to cnf. I know just how to convert cfg. What should I do with probabilities?

Example:

S->c W1 g:0.9|c W:0.1
W1->g W2 u:0.8|W2 u:0.2
W2->g W3 c:0.6|a W3 u:0.4
W3->g W4 c:0.5|W4 c:0.5
W4->a W5:0.7|c W5:0.3
W5->u W6:0.9|g W6:0.1
W6->a:0.6|u:0.4

I'm also looking for any information(perfectly some code examples) about Prediction Secondary RNA Structure. I would be happy for ane help!

Answer 1

Ok, I resolved my problem with CNF. If someone interested the SCFG in CNF:

S->C K:0.9|C W1:0.1
C->c:1.0
K->W1 L:1.0
L->g:1.0
W1->L J:0.8|J:0.2
J->W2 H:1.0
H->u:1.0
W2->L U:0.6|Y T:0.4
Y->a:1.0
T->W3 H:1.0
U->W3 O:1.0
O->c:1.0
W3->L R:0.5|W4 O:0.5
R->W4 O:1.0
W4->Y W5:0.7|O W5:0.3
W5->H W6:0.9|L W6:0.1
W6->a:0.6|u:0.4