boredbird / woe Goto Github PK

Hi, I am looking for documentation about your package : what can be done with it and how. Is there such a document somewhere? I have not been able to locate one so far. Thanks for your help.

克隆到本地安装后，运行HereWeGo.py程序报错。

check_point 函数在合并分段的时候有一个错误，
feature_process.py 112行这里：
pdf = df[(df[var] > split[i]) & (df[var] <= split[i+1])]
没有考虑到相邻的split中的sample数量都少于min_sample，但是合起来又大于min_sample的情况。
可以改为：

pre_left_position = float('-inf') # a value to remember the left point of the segmentation
for i in range(-1,split.__len__()-1):
    pdf = df[(df[var] > pre_left_position) & (df[var] <= split[i+1])]
    if (pdf.shape[0] < min_sample) or (np.unique(pdf['target']).__len__()<=1):
        #print(var, pre_left_position, i, "continue")
        continue
    else:
        new_split.append(split[i+1])
        pre_left_position = split[i+1]

离散型变量为什么要使用bin_data_split，因为bin_data_split中如果np.unique(var)<=8时候，只调用了下check_point()就直接retrun了，外面又调用check_point()，如果离散变量的个数>8怎么办？就按照百分位做了吗？

When I use proc_woe_continuous and eval_feature_detail to get iv results, there are too many msgs printed to terminal. Can it be quiet? Or can you use a debug parameter to control standard output?

Greetings and congrats for your work here.

I was wondering how can I make transformations on unseen data, I mean data for which we dont have a target variable. Is it possible? and then use the new woed features as input for other machine learning algorithm?

Thanks!

File "", line 8, in
import woe.config as config

File "apps/python/anaconda3/lib/python3.6/site-packages/woe/config.py", line 51
raise KeyError,"Invalid dtype specified! "
^
SyntaxError: invalid syntax

could not import woe.config as config
how to support python3 ?

Hi,

Is there anyway we can process pandas data frame instead of reading from csv file?

Thanks,
Bohan