OrdinalEncoder‘object没有属性'_missing_indices’

Question

我能够训练模型并成功地在Azure中部署端点。但是，当我测试端点时:它给出了下面的错误：

Failed to test real-time endpoint
{"status_code":400,"message":"'OrdinalEncoder' object has no attribute '_missing_indices'"}

在我的模型中，我使用了带ordinalEncoder的sklearn管道，运行在sklearn版本0.24.1上。守则的一部分如下：

numeric_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='most_frequent'))         
])

categorical_transformer = Pipeline(steps=[
    ('ordinal', OrdinalEncoder(handle_unknown='ignore'))
])

features_preprocessor = ColumnTransformer(
    transformers=[
        ('numeric',     numeric_transformer,     cols_tofill.values),
        ('categorical', categorical_transformer, ['Country','Status'])
    ], remainder='passthrough')

一直在尝试调试，但是如果我不知道这个错误意味着什么，那就很困难了。有人能告诉我吗？非常感谢你的帮助。

下面是我传递给Azure端点的' test‘选项卡的测试数据：

{  "data": [{"Country": "Japan",
             "Year": 2021,
             "Status": "Developed",
             "Adult_Mortality": 52,
             "Infant_Deaths": 1,
             "Alcohol": 1.9,
             "Percentage_Expenditure": 8000,
             "Hepatitis_B": 98,
             "Measles": 0,
             "BMI": 33,
             "Under-five_Deaths": 0,
             "Polio": 90,
             "Total_Expenditure": 5,
             "Diphtheria": 99,
             "HIV/AIDS": 0.1,
             "GDP": 57000,
             "Population": 6000000,
             "thinness_1-19_years": 2,
             "thinness_5-9_years": 2,
             "Income_composition_of_resources": 0.9,
             "Schooling": 15}],
   "method": "predict"
}

Answer 1

此Failed to test real-time endpoint {"status_code":400,"message":"'OrdinalEncoder' object has no attribute '_missing_indices'"}错误意味着您缺少一个或多个索引，您可以尝试以下方法来处理它：

 # stores the missing indices per category
        self._missing_indices = {}
        for cat_idx, categories_for_idx in enumerate(self.categories_):
            for i, cat in enumerate(categories_for_idx):
                if is_scalar_nan(cat):
                    self._missing_indices[cat_idx] = i
                    continue

         if np.dtype(self.dtype).kind != "f" and self._missing_indices:
            raise ValueError(
                "There are missing values in features "
                f"{list(self._missing_indices)}. For OrdinalEncoder to "
                "passthrough missing values, the dtype parameter must be a "
                "float"
            )

        return self

您可以参考encoders.py和sklearn.impute.MissingIndicator