1

我正在使用 Scikit-Learn 应用 SVM 算法来预测客户是否会选择房屋贷款。我希望将模型导出为 PMML 格式。数据集中的特征和标签如下所示：

特征
1. 访问频率
2. 对报价的响应
3. 网上银行设施
的使用 4. 储蓄账户
数量 5. 支票账户
数量 6. 开出的支票
数量 7. 数量完成的电子转帐
8. 获得的财产
9. 其他贷款行为
10. 收入

标签
是房屋贷款

模型生成正确，但无法导出到 PMML。代码粘贴在下面：

代码：

from sklearn.decomposition import PCA

from sklearn2pmml.decoration import ContinuousDomain

import pandas
import sklearn_pandas
from sklearn.svm import SVC



home_loan = pandas.read_csv('home-loan-dataset.csv')

home_loan = home_loan.drop(['CustID'], axis=1)

home_loan_df = pandas.concat((pandas.DataFrame(home_loan[:], columns = ['Frequencyofvisits','Responsetooffers','UsageofOnlineBankingFacility','Numberofsavingsaccount','Numberofcheckingaccount','Numberofcheckswritten','NumberofEFTsdone','PropertyAcquired','OtherLoansBehaviour','Income']), pandas.DataFrame(home_loan['IsHouseLoan'], columns = ["IsHouseLoan"])), axis = 1)

home_loan_mapper = sklearn_pandas.DataFrameMapper([
    (['Frequencyofvisits','Responsetooffers','UsageofOnlineBankingFacility','Numberofsavingsaccount','Numberofcheckingaccount','Numberofcheckswritten','NumberofEFTsdone','PropertyAcquired','OtherLoansBehaviour','Income'], [ContinuousDomain(), PCA(n_components = 3)]),
    ("IsHouseLoan", None)
])


home_loan = home_loan_df


home_loan_X = home_loan[['Frequencyofvisits','Responsetooffers','UsageofOnlineBankingFacility','Numberofsavingsaccount','Numberofcheckingaccount','Numberofcheckswritten','NumberofEFTsdone','PropertyAcquired','OtherLoansBehaviour','Income']]

home_loan_y = home_loan[['IsHouseLoan']]

# Classify using SVM

home_loan_classifier = SVC()



home_loan_classifier.fit(home_loan_X, home_loan_y.values.ravel())

SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,
    decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',
    max_iter=-1, probability=False, random_state=None, shrinking=True,
    tol=0.001, verbose=False)



#
# Conversion to PMML
#

from sklearn2pmml import sklearn2pmml

sklearn2pmml(home_loan_classifier, home_loan_mapper, "SVMHomeLoan.pmml", with_repr = True)


转换为 PMML 时显示以下错误：

错误：

C:\Python27\python.exe C:/Users/Admin/PycharmProjects/ML-Programs/Bank-Customer-Segmentation/svm-pmml.py
Aug 17, 2016 11:35:01 AM org.jpmml.sklearn.Main run
INFO: Parsing DataFrameMapper PKL..
Aug 17, 2016 11:35:01 AM org.jpmml.sklearn.Main run
INFO: Parsed DataFrameMapper PKL in 30 ms.
Aug 17, 2016 11:35:01 AM org.jpmml.sklearn.Main run
INFO: Converting DataFrameMapper..
Aug 17, 2016 11:35:01 AM org.jpmml.sklearn.Main run
SEVERE: Failed to convert DataFrameMapper
java.lang.IllegalArgumentException: The value of the sklearn2pmml.decoration.ContinuousDomain.data_min_ attribute (null) is not a supported array type
    at org.jpmml.sklearn.ClassDictUtil.getArray(ClassDictUtil.java:51)
    at sklearn2pmml.decoration.ContinuousDomain.getDataMin(ContinuousDomain.java:111)
    at sklearn2pmml.decoration.ContinuousDomain.encodeFeatures(ContinuousDomain.java:50)
    at sklearn_pandas.DataFrameMapper.encodeFeatures(DataFrameMapper.java:70)
    at org.jpmml.sklearn.Main.run(Main.java:146)
    at org.jpmml.sklearn.Main.main(Main.java:107)

Exception in thread "main" java.lang.IllegalArgumentException: The value of the sklearn2pmml.decoration.ContinuousDomain.data_min_ attribute (null) is not a supported array type
    at org.jpmml.sklearn.ClassDictUtil.getArray(ClassDictUtil.java:51)
    at sklearn2pmml.decoration.ContinuousDomain.getDataMin(ContinuousDomain.java:111)
    at sklearn2pmml.decoration.ContinuousDomain.encodeFeatures(ContinuousDomain.java:50)
    at sklearn_pandas.DataFrameMapper.encodeFeatures(DataFrameMapper.java:70)
    at org.jpmml.sklearn.Main.run(Main.java:146)
    at org.jpmml.sklearn.Main.main(Main.java:107)
Traceback (most recent call last):
  File "C:/Users/Admin/PycharmProjects/ML-Programs/Bank-Customer-Segmentation/svm-pmml.py", line 52, in <module>
    sklearn2pmml(home_loan_classifier, home_loan_mapper, "SVMHomeLoan.pmml", with_repr = True)
  File "C:\Python27\lib\site-packages\sklearn2pmml\__init__.py", line 56, in sklearn2pmml
    subprocess.check_call(cmd)
  File "C:\Python27\lib\subprocess.py", line 540, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['java', '-cp', 'C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\guava-19.0.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\istack-commons-runtime-2.21.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\jaxb-core-2.2.11.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\jaxb-runtime-2.2.11.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\jcommander-1.48.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\jpmml-converter-1.0.7.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\jpmml-sklearn-1.0-SNAPSHOT.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\jpmml-xgboost-1.0.5.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\pmml-agent-1.2.16.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\pmml-model-1.2.16.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\pmml-model-metro-1.2.16.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\pmml-schema-1.2.16.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\pyrolite-4.12.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\serpent-1.12.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\slf4j-api-1.7.21.jar;C:\\Python27\\lib\\site-packages\\sklearn2pmml\\resources\\slf4j-jdk14-1.7.21.jar', 'org.jpmml.sklearn.Main', '--pkl-estimator-input', 'c:\\users\\Admin\\appdata\\local\\temp\\tmplgmrjq.pkl', '--repr-estimator', "SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n  decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',\n  max_iter=-1, probability=False, random_state=None, shrinking=True,\n  tol=0.001, verbose=False)", '--pkl-mapper-input', 'c:\\users\\Admin\\appdata\\local\\temp\\tmpobahse.pkl', '--repr-mapper', "DataFrameMapper(features=[(['Frequencyofvisits', 'Responsetooffers', 'UsageofOnlineBankingFacility', 'Numberofsavingsaccount', 'Numberofcheckingaccount', 'Numberofcheckswritten', 'NumberofEFTsdone', 'PropertyAcquired', 'OtherLoansBehavior', 'Income100000'], TransformerPipeline(steps=[('continuousdomain', ContinuousDomain(invalid_value_treatment='return_invalid')), ('pca', PCA(copy=True, n_components=3, whiten=False))])), ('IsHouseLoan', None)],\n        sparse=False)", '--pmml-output', 'SVMHomeLoan.pmml']' returned non-zero exit status 1


可能是什么原因 ？

4

1 回答 1

2

显然，您的数据列之一不符合sklearn2pmml.decoration.ContinuousDomain转换的预期。在没有看到您的数据的情况下，不可能说出哪一列以及问题的确切性质是什么（例如，分类操作类型而不是连续的、错误的数字数据类型、列包含 NA 值等）。

您在这里有两个选择：

  1. 识别行为异常的列，并修复数据问题，以便ContinuousDomain转换正常工作。
  2. ContinuousDomain从列表转换列表中排除。

目前，您正在使用直接从 sklearn2pmml README.md 文件复制的数据预处理逻辑。请重新处理它以匹配您的数据 -[ContinuousDomain(), PCA(n_components = 3)]转换不太可能是您的用例的正确解决方案。

此外，这个问题是相当特定于sklearn2pmml包的。如果您在 sklearn2pmml 问题跟踪器中打开问题，您可能会得到更好/更快的回复。

于 2016-08-17T06:42:18.553 回答