xiaolang

[email protected]

Zhejiang University

Results 4 comments of


                                            xiaolang

hash trick

只要特征取值少于10的五次方就行

I found that your code was wrong in 100 and 101 lines in the gen_samples.py file

@YSZYCF 这两个list在下面代码都有append操作的，你注意看下

I found that your code was wrong in 100 and 101 lines in the gen_samples.py file

@YSZYCF 是我疏忽了，漏了两行代码，现已加上，感谢指正。

res['pred']*1.2 + res['prob']?

@YSZYCF 为了融合两个模型的结果融合比例1.2是从线下验证集中得到的