为什么我在 python 中创建的类无法识别设置为 False 的属性？

Question

我正在尝试在房价上竞争 - 高级回归技术 Kaggle 比赛

我正在编写一个自定义转换器，它可以与添加组合属性的 Scikit-Learn 功能无缝协作

转换器有四个超参数 (add_baths, add_bsmt_baths, add_above_grade_baths, add_porch_area) 默认设置为 True。这个超参数可以让我轻松找出添加这个属性是否有助于机器学习算法。

但问题是，当我将这些超参数之一设置为 False 时，该类仍会返回该列，就像我将其设置为 True 一样

class CombinedAttributesAdder(BaseEstimator, TransformerMixin):
        def __init__(self, add_baths=True, add_bsmt_baths=True, add_above_grade_baths=True, add_porch_area=True):
            self.add_baths = add_baths
            self.add_bsmt_baths = add_bsmt_baths
            self.add_above_grade_baths = add_above_grade_baths
            self.add_porch_area = add_porch_area
        def fit(self, X, y=None):
            return self
        def transform(self, X):
            X['T_FlrSF'] = X['1stFlrSF'] + X['2ndFlrSF']
            if self.add_baths:
                X['T_Bath'] = X['BsmtFullBath'] + X['BsmtHalfBath'] + X['FullBath'] + X['HalfBath']
            if self.add_bsmt_baths:
                X['T_BsmtBath'] = X['BsmtFullBath'] + X['BsmtHalfBath']
            if self.add_above_grade_baths:
                X['T_agBath'] = X['FullBath'] + X['HalfBath']
            if self.add_porch_area:
                X['T_Porch'] = X['OpenPorchSF'] + X['EnclosedPorch'] + X['3SsnPorch'] + X['ScreenPorch']
            return X

    attr_adder = CombinedAttributesAdder(add_baths=False, add_bsmt_baths=False)
    housing_extra_attribs = attr_adder.transform(housing)

这里应该返回所有列，因为我设置了参数add_baths=False, add_bsmt_baths=False 它不应该创建T_Bath 也不应该创建T_BsmtBath 列

housing_extra_attribs.columns

...

 Index(['Id', 'MSSubClass', 'MSZoning', 'LotFrontage', 'LotArea', 'Street',
   'Alley', 'LotShape', 'LandContour', 'Utilities', 'LotConfig',
   'LandSlope', 'Neighborhood', 'Condition1', 'Condition2', 'BldgType',
   'HouseStyle', 'OverallQual', 'OverallCond', 'YearBuilt', 'YearRemodAdd',
   'RoofStyle', 'RoofMatl', 'Exterior1st', 'Exterior2nd', 'MasVnrType',
   'MasVnrArea', 'ExterQual', 'ExterCond', 'Foundation', 'BsmtQual',
   'BsmtCond', 'BsmtExposure', 'BsmtFinType1', 'BsmtFinSF1',
   'BsmtFinType2', 'BsmtFinSF2', 'BsmtUnfSF', 'TotalBsmtSF', 'Heating',
   'HeatingQC', 'CentralAir', 'Electrical', '1stFlrSF', '2ndFlrSF',
   'LowQualFinSF', 'GrLivArea', 'BsmtFullBath', 'BsmtHalfBath', 'FullBath',
   'HalfBath', 'BedroomAbvGr', 'KitchenAbvGr', 'KitchenQual',
   'TotRmsAbvGrd', 'Functional', 'Fireplaces', 'FireplaceQu', 'GarageType',
   'GarageYrBlt', 'GarageFinish', 'GarageCars', 'GarageArea', 'GarageQual',
   'GarageCond', 'PavedDrive', 'WoodDeckSF', 'OpenPorchSF',
   'EnclosedPorch', '3SsnPorch', 'ScreenPorch', 'PoolArea', 'PoolQC',
   'MiscFeature', 'MiscVal', 'MoSold', 'YrSold', 'SaleType',
   'SaleCondition', 'T_FlrSF', 'T_Bath', 'T_BsmtBath', 'T_agBath',
   'T_Porch'],
  dtype='object')

Answer 1

如果不复制转换部分中的训练数据，您正在修改原始训练数据，因此如果您之前已经使用 True 属性对其进行了转换，那么这些列已经在之前创建。 请尝试以下方法，但请确保之前没有修改过训练数据：

class CombinedAttributesAdder(BaseEstimator, TransformerMixin):
        def __init__(self, add_baths=True, add_bsmt_baths=True, add_above_grade_baths=True, add_porch_area=True):
            self.add_baths = add_baths
            self.add_bsmt_baths = add_bsmt_baths
            self.add_above_grade_baths = add_above_grade_baths
            self.add_porch_area = add_porch_area
        def fit(self, X, y=None):
            return self
        def transform(self, X):
            X_=X.copy()
            X_['T_FlrSF'] = X_['1stFlrSF'] + X_['2ndFlrSF']
            if self.add_baths:
                X_['T_Bath'] = X_['BsmtFullBath'] + X_['BsmtHalfBath'] + X_['FullBath'] + X_['HalfBath']
            if self.add_bsmt_baths:
                X_['T_BsmtBath'] = X_['BsmtFullBath'] + X_['BsmtHalfBath']
            if self.add_above_grade_baths:
                X_['T_agBath'] = X_['FullBath'] + X_['HalfBath']
            if self.add_porch_area:
                X_['T_Porch'] = X_['OpenPorchSF'] + X_['EnclosedPorch'] + X_['3SsnPorch'] + X_['ScreenPorch']
            return X_