https://github.com/duanhongyi/genius
a chinese segment base on crf
https://github.com/duanhongyi/genius
Last synced: 2 months ago
JSON representation
a chinese segment base on crf
- Host: GitHub
- URL: https://github.com/duanhongyi/genius
- Owner: duanhongyi
- License: other
- Created: 2013-08-20T06:54:42.000Z (over 12 years ago)
- Default Branch: master
- Last Pushed: 2018-12-19T16:03:58.000Z (about 7 years ago)
- Last Synced: 2024-11-10T10:51:54.598Z (over 1 year ago)
- Language: Python
- Size: 96.8 MB
- Stars: 234
- Watchers: 26
- Forks: 65
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGES.txt
Awesome Lists containing this project
- awesome-machine-master - genius - A Chinese segment base on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-machine-learning - genius - A Chinese segment based on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-machine-learning - genius - A Chinese segment base on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-machine-learning - genius - A Chinese segment based on Conditional Random Field. (Python / General-Purpose Machine Learning)
- fucking-awesome-machine-learning - genius - A Chinese segment based on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-machine-learning - genius - A Chinese segment based on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-machine-learning - genius - A Chinese segment base on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-machine-learning-cn - 官网
- awesome-chinese-nlp - Genius 中文分词
- awesome-machine-learning - genius - A Chinese segment based on Conditional Random Field. (Python / General-Purpose Machine Learning)
- awesome-advanced-metering-infrastructure - genius - A Chinese segment base on Conditional Random Field. (Python / General-Purpose Machine Learning)
README
Genius
========
Genius是一个开源的python中文分词组件，采用 CRF(Conditional Random Field)条件随机场算法。
Feature
========
* 支持python2.x、python3.x以及pypy2.x。
* 支持简单的pinyin分词
* 支持用户自定义break
* 支持用户自定义合并词典
* 支持词性标注
Source Install
==========
* 安装git: 1) ubuntu or debian `apt-get install git` 2) fedora or redhat `yum install git`
* 下载代码：`git clone https://github.com/duanhongyi/genius.git`
* 安装代码：`python setup.py install`
Pypi Install
============
* 执行命令：`easy_install genius`或者`pip install genius`
Algorithm
==========
* 采用trie树进行合并词典查找
* 基于wapiti实现条件随机场分词
* 可以通过genius.loader.ResourceLoader来重载默认的字典
功能 1)：分词`genius.seg_text`方法
==============
* `genius.seg_text`函数接受5个参数，其中text是必填参数:
* `text`第一个参数为需要分词的字符
* `use_break`代表对分词结构进行打断处理，默认值`True`
* `use_combine`代表是否使用字典进行词合并，默认值`False`
* `use_tagging`代表是否进行词性标注，默认值`True`
* `use_pinyin_segment`代表是否对拼音进行分词处理，默认值`True`
代码示例( 全功能分词 )
#encoding=utf-8
import genius
text = u"""昨天,我和施瓦布先生一起与部分企业家进行了交流,大家对中国经济当前、未来发展的态势、走势都十分关心。"""
seg_list = genius.seg_text(
text,
use_combine=True,
use_pinyin_segment=True,
use_tagging=True,
use_break=True
)
print('\n'.join(['%s\t%s' % (word.text, word.tagging) for word in seg_list]))
功能 2)：面向索引分词
==============
* `genius.seg_keywords`方法专门为搜索引擎索引准备，保留歧义分割，其中text是必填参数。
* `text`第一个参数为需要分词的字符
* `use_break`代表对分词结构进行打断处理，默认值`True`
* `use_tagging`代表是否进行词性标注，默认值`False`
* `use_pinyin_segment`代表是否对拼音进行分词处理，默认值`False`
* 由于合并操作与此方法有意义上的冲突，此方法并不提供合并功能；并且如果采用此方法做索引时候，检索时不推荐`genius.seg_text`使用`use_combine=True`参数。
代码示例
#encoding=utf-8
import genius
seg_list = genius.seg_keywords(u'南京市长江大桥')
print('\n'.join([word.text for word in seg_list]))
功能 3)：关键词提取
==============
* `genius.extract_tag`方法专门为提取tag关键字准备，其中text是必填参数。
* `text`第一个参数为需要分词的字符
* `use_break`代表对分词结构进行打断处理，默认值`True`
* `use_combine`代表是否使用字典进行词合并，默认值`False`
* `use_pinyin_segment`代表是否对拼音进行分词处理，默认值`False`
代码示例
#encoding=utf-8
import genius
tag_list = genius.extract_tag(u'南京市长江大桥')
print('\n'.join(tag_list))
其他说明 4)：
=================
* 目前分词语料出自人民日报1998年1月份，所以对于新闻类文章分词较为准确。
* CRF分词效果很大程度上依赖于训练语料的类别以及覆盖度，若解决语料问题分词和标注效果还有很大的提升空间。