条件随机场java实现_中文分词模型之条件随机场(CRFs)模型

对于CRFs模型，一直没有找到比较直观简单的Java实现。没有办法，就参考了博客：

该博客的细节非常清楚。在windows系统中使用CRF++，可以下载编译好的程序，地址如下：

最好还是能自己找到原始地址去，推荐一个过墙的东东:http://honx.in/i/VKEHUuz5NEy3Un5i ，我一直在用，感觉还行。

最后直接给出用pku的数据集在backoff测试的结果：

封闭测试：(用训练集测试)

=== SUMMARY:

=== TOTAL INSERTIONS:987

=== TOTAL DELETIONS:1312

=== TOTAL SUBSTITUTIONS:2277

=== TOTAL NCHANGE:4576

=== TOTAL TRUE WORD COUNT:1109947

=== TOTAL TEST WORD COUNT:1109622

=== TOTAL TRUE WORDS RECALL:0.997

=== TOTAL TEST WORDS PRECISION:0.997

=== F MEASURE:0.997

=== OOV Rate:0.000

=== OOV Recall Rate:--

=== IV Recall Rate:0.997

###pku_crf_training.word.utf8987131222774576110994711096220.9970.9970.9970.000--0.997

开放测试：(用gold/test测试)

=== SUMMARY:

=== TOTAL INSERTIONS:1492

=== TOTAL DELETIONS:3150

=== TOTAL SUBSTITUTIONS:4966

=== TOTAL NCHANGE:9608

=== TOTAL TRUE WORD COUNT:104372

=== TOTAL TEST WORD COUNT:102714

=== TOTAL TRUE WORDS RECALL:0.922

=== TOTAL TEST WORDS PRECISION:0.937

=== F MEASURE:0.930

=== OOV Rate:0.058

=== OOV Recall Rate:0.562

=== IV Recall Rate:0.944

###pku_crf_test.word.utf814923150496696081043721027140.9220.9370.9300.0580.5620.944

开放测试也有93.7%的正确率，这还是没有经过调优。都是CRF++默认的的特征模板。