Scikit learn 无法创建PMML，因为未指定输入要素的数量

我无法将以下管道转换为pmml，因为“未指定输入功能的数量”

再现错误的最小示例管道为：

import pandas as pd
from sklearn.compose import ColumnTransformer
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn2pmml import sklearn2pmml
from sklearn2pmml.pipeline import PMMLPipeline

if __name__ == '__main__':

    data_dict = {
        'age': [1, 2, 3],
        'day_of_week': ['monday', 'tuesday', 'wednesday'],
        'y': [5, 6, 7]
    }

    data = pd.DataFrame(data_dict, columns=data_dict)

    numeric_features = ['age']
    numeric_transformer = Pipeline(steps=[
        ('scaler', StandardScaler())])

    categorical_features = ['day_of_week']
    categorical_transformer = Pipeline(steps=[
        ('onehot', OneHotEncoder(handle_unknown='ignore', categories='auto'))])

    preprocessor = ColumnTransformer(
        transformers=[
            ('numerical', numeric_transformer, numeric_features),
            ('categorical', categorical_transformer, categorical_features)])

    pipeline = Pipeline(
        steps=[
            ('preprocessor', preprocessor),
            ('classifier', RandomForestRegressor(n_estimators=60))])

    X = data.drop(labels=['y'], axis=1)
    y = data['y']

    X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.8, random_state=30)

    trained_model = pipeline.fit(X=X_train, y=y_train)

    pmml_pipeline = PMMLPipeline([
        ("pipeline", pipeline)
    ])

    sklearn2pmml(pipeline=pmml_pipeline, pmml='RandomForestRegressor2.pmml', with_repr=True)

我从

sklearn2pmml

获得的Java错误消息是：

Standard output is empty
Standard error:
May 25, 2020 9:37:56 AM org.jpmml.sklearn.Main run
INFO: Parsing PKL..
May 25, 2020 9:38:07 AM org.jpmml.sklearn.Main run
INFO: Parsed PKL in 11453 ms.
May 25, 2020 9:38:07 AM org.jpmml.sklearn.Main run
INFO: Converting..
May 25, 2020 9:38:07 AM sklearn2pmml.pipeline.PMMLPipeline initTargetFields
WARNING: Attribute 'sklearn2pmml.pipeline.PMMLPipeline.target_fields' is not set. Assuming y as the name of the target field
May 25, 2020 9:38:07 AM org.jpmml.sklearn.Main run
SEVERE: Failed to convert
java.lang.IllegalArgumentException: The transformer object of the first step (Python class sklearn.pipeline.Pipeline) does not specify the number of input features
    at sklearn2pmml.pipeline.PMMLPipeline.initActiveFields(PMMLPipeline.java:522)
    at sklearn2pmml.pipeline.PMMLPipeline.encodePMML(PMMLPipeline.java:214)
    at org.jpmml.sklearn.Main.run(Main.java:228)
    at org.jpmml.sklearn.Main.main(Main.java:148)

Exception in thread "main" java.lang.IllegalArgumentException: The transformer object of the first step (Python class sklearn.pipeline.Pipeline) does not specify the number of input features
    at sklearn2pmml.pipeline.PMMLPipeline.initActiveFields(PMMLPipeline.java:522)
    at sklearn2pmml.pipeline.PMMLPipeline.encodePMML(PMMLPipeline.java:214)
    at org.jpmml.sklearn.Main.run(Main.java:228)
    at org.jpmml.sklearn.Main.main(Main.java:148)


Process finished with exit code 1

---

基于

sklearn.pipeline.pipeline

创建单个步骤

sklearn2pmml.pipline.PMMLPipeline

有什么意义

省略此no op，则转换应成功：

pipeline = PMMLPipeline(
    steps=[
        ('preprocessor', preprocessor),
        ('classifier', RandomForestRegressor(n_estimators=60))])
pipeline.fit(X=X_train, y=y_train)
sklearn2pmml(pipeline, "RandomForestRegressor2.pmml")

---

你说得对。澄清：

PMMLPipeline

和

Pipeline

能否互换使用？因此，如果我想导出为PMML，我将使用

PMMLPipeline

否则

Pipeline

？

PMMLPipeline

是

Pipeline

的子类，因此在所有情况下都可以使用它作为替代品。

PMMLPipeline

的目标是提供Scikit Learn中不可用的额外功能（例如捕获特征元数据、执行决策工程和模型验证）。最初的错误是关于不必要的管道嵌套。当然，SkLearn2PMML包可以检测并打开这些。