-
Notifications
You must be signed in to change notification settings - Fork 16
Supported MeCab Options
Brooke M. Fujita edited this page Feb 9, 2015
·
1 revision
The MeCab parsing and output from natto can be customized by using the following options:
- rcfile -- resource file
- dicdir -- system dicdir
- userdic -- user dictionary
- lattice-level -- lattice information level (DEPRECATED)
- output-format-type -- output format type (wakati, chasen, yomi, etc.)
- all-morphs -- output all morphs (default false)
- nbest -- output N best results (integer, default 1), requires lattice level >= 1
- partial -- partial parsing mode (default false)
- marginal -- output marginal probability (default false)
- max-grouping-size -- maximum grouping size for unknown words (integer, default 24)
- node-format -- user-defined node format
- unk-format -- user-defined unknown node format
- bos-format -- user-defined beginning-of-sentence format
- eos-format -- user-defined end-of-sentence format
- eon-format -- user-defined end-of-NBest format
- unk-feature -- feature for unknown morpheme
- input-buffer-size -- set input buffer size (default 8192)
- allocate-sentence -- allocate new memory for input sentence
- theta -- temperature parameter theta (float, default 0.75)
- cost-factor -- cost factor (integer, default 700)