如何修复使用make_pipeline时的TypeError: too many positional arguments错误

原代码 #

from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
from sklearn.pipeline import make_pipeline
from sklearn.model_selection import validation_curve

def PolynomialRegression(degree=2, **kwargs):
     return make_pipeline(PolynomialFeatures(degree), LinearRegression(**kwargs))

train_score, val_score = validation_curve(PolynomialRegression(), X, y,
                                      'polynomialfeatures__degree',
                                      degree, cv=7)

报错信息 #

File ~/opt/miniconda3/lib/python3.10/inspect.py:3108, in Signature._bind(self, args, kwargs, partial)
3104 else:
3105 if param.kind in (_VAR_KEYWORD, _KEYWORD_ONLY):
3106 # Looks like we have no parameter for this positional
3107 # argument
-> 3108 raise TypeError(
3109 'too many positional arguments') from None
3111 if param.kind == _VAR_POSITIONAL:
3112 # We have an '*args'-like argument, let's fill it with
3113 # all positional arguments we have left and move on to
3114 # the next phase
3115 values = [arg_val]

TypeError: too many positional arguments

问题分析与修复 #

错误根源在于validation_curve的参数传递逻辑：

1. 模型参数错误：validation_curve要求第一个参数是模型构造函数（而非已实例化的模型对象），你传入的PolynomialRegression()是已创建好的模型实例，导致函数无法动态生成不同参数的模型。
2. 参数值未定义：代码中degree变量未明确赋值，需要传入一个包含待测试多项式度数的列表。

修正后的代码：

from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
from sklearn.pipeline import make_pipeline
from sklearn.model_selection import validation_curve

def PolynomialRegression(degree=2, **kwargs):
     return make_pipeline(PolynomialFeatures(degree), LinearRegression(**kwargs))

# 定义要测试的多项式度数范围
degrees = [1, 2, 3, 4, 5]
train_score, val_score = validation_curve(PolynomialRegression, X, y,
                                      'polynomialfeatures__degree',
                                      degrees, cv=7)

关键修正点 #

• 将PolynomialRegression()改为PolynomialRegression：让validation_curve可以根据不同的degree参数动态实例化模型。
• 新增degrees列表：明确指定要验证的多项式度数，替换未定义的degree变量。

内容的提问来源于stack exchange，提问作者zjohn