xgboost : TypeError: predict() got an unexpected keyword argument 'pred_contribs'

Posted 2023-03-12

【中文标题】xgboost : TypeError: predict() got an unexpected keyword argument \'pred_contribs\'

【英文标题】：xgboost : TypeError: predict() got an unexpected keyword argument 'pred_contribs'xgboost : TypeError: predict() got an unexpected keyword argument 'pred_contribs'

【发布时间】：2022-01-12 06:14:30

【问题描述】：

我使用 xgboost v0.90 训练了一个模型，使其与 AWS SageMaker ML Engine 兼容。我正在做通常的编码和超参数调整。下面是一些代码：

import pandas as pd
import pickle
from xgboost import XGBRegressor
from sklearn.model_selection import train_test_split, GridSearchCV, RandomizedSearchCV
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder 


# split df into train and test
X_train, X_test, y_train, y_test = train_test_split(df.iloc[:,0:21], df.iloc[:,-1], test_size=0.1)

X_train.shape
(1000,21)

# Encode categorical variables  
cat_vars = ['cat1','cat2','cat3']
cat_transform = ColumnTransformer([('cat', OneHotEncoder(handle_unknown='ignore'), cat_vars)], remainder='passthrough')

encoder = cat_transform.fit(X_train)
X_train = encoder.transform(X_train)
X_test = encoder.transform(X_test)

X_train.shape
(1000,420)

# Define a xgboost regression model
model = XGBRegressor()

# Do hyper-parameter tuning
.....

# Fit model
model.fit(X_train, y_train)


# Forecast on test data
y_pred = model.predict(X_test, pred_contribs=True)
y_pred

我已经安装了 SHAP 并根据文档 [1]，.predict() 采用 pred_contribs 参数。追溯：

    TypeError                                 Traceback (most recent call last)
<ipython-input-1119-37e607e853fd> in <module>
      1 # Forecast on test data
----> 2 y_pred = model.predict(X_test, pred_contribs=True)
      3 y_pred

TypeError: predict() got an unexpected keyword argument 'pred_contribs'

[1]https://shap.readthedocs.io/en/latest/example_notebooks/tabular_examples/tree_based_models/Python%20Version%20of%20Tree%20SHAP.html?highlight=pred_contribs

[2]https://xgboost.readthedocs.io/en/latest/python/python_api.html#xgboost.Booster.predict

【参考方案1】：

XGBRegressor 的 predict 没有pred_contribs 参数，它只有这些参数：

predict(self, X, output_margin=False, ntree_limit=None,
        validate_features=True, base_margin=None, iteration_range=None)

你说的参数pred_contribs在xgboost.Booster.predict()中，这是一个低级XGBoost，它有这些参数：

predict(self, data: xgboost.core.DMatrix, output_margin: bool = False,
        ntree_limit: int = 0, pred_leaf: bool = False, pred_contribs: bool = False,
        approx_contribs: bool = False, pred_interactions: bool = False,
        validate_features: bool = True, training: bool = False,
        iteration_range: Tuple[int, int] = (0, 0), strict_shape: bool = False)

所以你必须使用它：

data =  xgboost.DMatrix(X, label = y)
model = xgboost.train("learning_rate": 0.01, "max_depth": 4, data)
model.predict(data, pred_contribs = True)

【讨论】：

我明白了。什么是低级 XGBoost？

XGBRegressor 和 XGBClassifier 是类似 sklearn 的包装器，可以使用 XGBRegressor 和 XGBClassifier 完成的所有操作都可以通过底层的 xgboost.train 函数完成。反过来，这显然不是真的。 xgboost.train 有更多参数，它让您可以更好地控制训练、验证和预测。这就是为什么xgboost.train 有pred_contribs 参数而XGBRegressor 没有。阅读this Stack Overflow 答案以获取更多信息。

正确。该模型是使用 XGBRegressor() 训练的，我无法更新它。当我进行推理/预测时会发生异常。我想我需要一个中间人将测试数据更新为 xgboost.DMatrix 然后调用 predict 方法？

你必须使用xgboost.train()训练你的模型