Scroll to navigation

TRAIN-KYTEA(1) General Commands Manual TRAIN-KYTEA(1)

NAME

kytea — a word segmentation/pronunciation estimation tool

SYNOPSIS

train-kytea [options]

DESCRIPTION

This manual page documents briefly the train-kytea command.

This manual page was written for the Debian distribution because the original program does not have a manual page. Instead, it has documentation in the GNU Info format; see below.

kytea is morphological analysis system based on pointwise predictors. It separetes sentences into words, tagging and predict pronunciations. The pronunciation of KyTea is same as cutie.

OPTIONS

A summary of options is included below.

Input/Output Options:

The text encoding to be used (utf8/euc/sjis; default: utf8)
A fully annotated training corpus (multiple possible)
A training corpus that is tokenized with no tags (multiple possible)
A partially annotated training corpus (multiple possible)
A confidence annotated training corpus (multiple possible)
A file containing features generated by -featout
A dictionary file (one 'word/pron' entry per line, multiple possible)
A file of subword units. This will enable unknown word PE.
The file to write the trained model to
Print a text model (instead of the default binary)
Write the features used in training the model to this file

Model Training Options (basic)

Don't train a word segmentation model
Skip the training of tagging, do only word segmentation
Train the nth tag with a global model (good for POS, bad for PE)
The debugging level during training (0=silent, 1=normal, 2=detailed)

Model Training Options (for advanced users):

The character window to use for WS (3)
The character n-gram length to use for WS for WS (3)
The character type window to use for WS (3)
The character type n-gram length to use for WS for WS (3)
Dictionary words greater than -dictn will be grouped together (4)
Language model n-gram order for unknown words (3)
The epsilon stopping criterion for classifier training
The cost hyperparameter for classifier training
Don't use a bias value in classifier training
The solver (1=SVM, 7=logistic regression, etc.; default 1, see LIBLINEAR documentation for more details)

Format Options (for advanced users):

The separator for words in full annotation (" ")
The separator for tags in full/partial annotation ("/")
The separator for candidates in full/partial annotation ("&")
Indicates unannotated boundaries in partial annotation (" ")
Indicates skipped boundaries in partial annotation ("?")
Indicates non-existence of boundaries in partial annotation ("-")
Indicates existence of boundaries in partial annotation ("|")

AUTHOR

This manual page was written by Koichi Akabe vbkaisetsu@gmail.com for the Debian system (and may be used by others). Permission is granted to copy, distribute and/or modify this document under the terms of the GNU General Public License, Version 2 any later version published by the Free Software Foundation.

On Debian systems, the complete text of the GNU General Public License can be found in /usr/share/common-licenses/GPL.