教师和学生相似度是1但老师和教师相似度小于0.9

[funicia]

现在行为

synonyms.compare('教育机构学生', '教育机构教师')
结果是1.0 ，这两个字符串相似度应该低于0更合理些。

synonyms.compare('学生', '教师')
这个结果也是1.0

synonyms.compare('学生', '老师')
这个结果是0.896

synonyms.compare('教师', '老师')
这个结果是0.898

操作系统

• Windows
• Linux(Debian, CentOS, Ubuntu, etc.)

解决方案

代码版本

Open Source for the World

[hailiang-wang]

I can not fix this badcase at the moment.

Thanks for your feedback, the badcases can not be fixed with a big corpus, as it is done in machine learning way.
The model pretrained in Synonyms uses public corpus four years ago, so, the open data for training should be available.

To learn more about how Synonyms works, please read this book.
https://github.com/l11x0m7/book-of-qna-code