无法使用AIF360调用DataFrame对象

Question

我遇到了TypeError：'DataFrame‘对象不能用下面的代码调用。有人能帮上忙吗？谢谢。

%cd -
dataset_orig = df_data_1(protected_attribute_names=['Gender'],
privileged_classes=['Male'],      
features_to_drop=[]) 

dataset_orig_train, dataset_orig_test = dataset_orig.split([0.7], shuffle=True)

privileged_groups = [{'Gender': 1}]
unprivileged_groups = [{'Gender': 0}]

/home/wsuser/work

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-59-8c624cfec261> in <module>
      5 # consider in this evaluation
      6 privileged_classes=['Male'],      # male is considered privileged
----> 7 features_to_drop=[]) # ignore all other attributes
      8 
      9 dataset_orig_train, dataset_orig_test = dataset_orig.split([0.7], shuffle=True)

TypeError: 'DataFrame' object is not callable

Answer 1

看起来df_data_1就是你的数据集数据框架，对吧？如果是，则需要更新脚本以将其转换为StandardDataset

from aif360.datasets import StandardDataset

dataset_orig = StandardDataset(df_data_1, 
                               protected_attribute_names=['Gender'],
                               privileged_classes=['Male'],
                               features_to_drop=[],
                               favorable_classes=[1]  # Update this with label values which are considered favorable in your dataset
                              ) 

我不确定您的数据集是什么样子，但您可以调整完整的可重现示例here来为您的数据集执行此过程。