pycorrector icon indicating copy to clipboard operation
pycorrector copied to clipboard

MacBERT等深度模型误纠解决思路

Open ZTurboX opened this issue 3 years ago • 4 comments

MacBERT等深度模型结合ngram规则纠错效果相对很好,但深度模型误纠比较高,请问有什么解决思路呢

ZTurboX avatar May 24 '22 00:05 ZTurboX

1、模型优化:补充负例case(无错样本),把误纠的填进去; 2、专名过滤:人名、地名、专名等词加到 confusion dict,过滤处理; 3、输出macbert纠错置信度,只纠正高置信度错误。

shibing624 avatar May 24 '22 02:05 shibing624

1、模型优化:补充负例case(无错样本),把误纠的填进去; 2、专名过滤:人名、地名、专名等词加到 confusion dict,过滤处理; 3、输出macbert纠错置信度,只纠正高置信度错误。

macbert纠错置信度是怎么算的

ZTurboX avatar May 24 '22 04:05 ZTurboX

softmax的概率值就是

shibing624 avatar May 24 '22 06:05 shibing624

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.(由于长期不活动,机器人自动关闭此问题,如果需要欢迎提问)

stale[bot] avatar Aug 13 '22 08:08 stale[bot]