issues
search
amphibian-dev
/
toad
ESC Team's credit scorecard tools.
https://toad.readthedocs.io
MIT License
474
stars
173
forks
source link
200万数据,500维特征,卡方分箱很慢,有没有好办法?
#111
Open
dinglei8908
opened
1 year ago
dinglei8908
commented
1 year ago
RT
Secbone
commented
1 year ago
@dinglei8908 如果可以接受误差的话,可以尝试先等频分成m箱(如1000),然后再使用卡方分箱分成n箱,n<<m
RT