You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -169,7 +169,7 @@ Note that this implementation does not apply innovations such as [Joop Leo's imp
169
169
For a faster parser that work on non-probabilistic grammars, look into [nearley](nearley.js.org).
170
170
171
171
### Limitations
172
-
* I have not provisioned for ε-rules
172
+
* I have not provisioned for ε-rules (rules with an empty right hand side)
173
173
* Rule probability estimation may be performed using the inside-outside algorithm, but is not currently implemented
174
174
* Higher level concepts such as wildcards, * and + are not implemented
175
175
* Viterbi parsing (querying the most likely parse tree) only returns one single parse. In the case of an ambiguous sentence in which multiple dervation have the highest probability, the returned parse is not guaranteed the left-most parse (I think).
0 commit comments