HillZhang1999 / NaSGEC

Code & Data for our Paper "NaSGEC: Multi-Domain Chinese Grammatical Error Correction for Native Speaker Texts" (ACL 2023 Findings)
https://arxiv.org/abs/2305.16023
75 stars 6 forks source link

Indicators of Domain Shifts #4

Open GMago-LeWay opened 1 year ago

GMago-LeWay commented 1 year ago

Hi, I noticed that the section 5.2 in NaSGEC paper described three indicators of domain shifts: VO, TDS, EPO. I try to reproduce it but I meet some trouble in TDS, EPO. Is there any related code? Or could you please provide some code for this part? Thanks.

HillZhang1999 commented 1 year ago

Hi, may I ask which problem did you meet? I may give you some advice. I will also consider open-sourcing code related to the calculation of these indicators later this month.

GMago-LeWay commented 1 year ago

In Error Pattern Overlap (EPO), what does the error pattern mean? Is it error type+correct content like |W|||爱 死|, |M|||在|?

HillZhang1999 commented 1 year ago

Actually, we define an error pattern as just a tuple of erroneous span and correction, which is: <蓝球, 篮球>.