我正在尝试在pipeline(StandardScaler, DecisionTreeClassifier)上运行k折交叉验证。

首先，我导入数据。

data = pd.read_csv('train_strokes.csv')

然后预处理数据框

# 预处理数据
data.drop('id', axis=1, inplace=True)
data['age'] = data['age'].apply(lambda x: x if round(x) else np.nan)
data['bmi'] = data['bmi'].apply(lambda bmi: bmi if 12 < bmi < 45 else np.nan)
data['gender'] = data['gender'].apply(lambda gender: gender if gender == 'Female' or gender == 'Male' else np.nan)
data.sort_values(['gender', 'age', 'bmi'], inplace=True)
data['bmi'].ffill(inplace=True)
data.dropna(axis=0, inplace=True)
data.reset_index(drop=True, inplace=True)
# 将分类数据转换为数值
enc = LabelEncoder()
data['gender'] = enc.fit_transform(data['gender'])
data['work_type'] = enc.fit_transform(data['work_type'])
data['Residence_type'] = enc.fit_transform(data['Residence_type'])
data['smoking_status'] = enc.fit_transform(data['smoking_status'])
data['ever_married'] = enc.fit_transform(data['ever_married'])

然后分割特征和目标

target = data['stroke']
feat = data.drop('stroke', axis=1)

并使用SMOTE来平衡数据

sm = SMOTE(random_state=1)
feat, target = sm.fit_resample(feat, target)
feat['age'] = feat['age'].apply(lambda x: round(x))
feat['hypertension'] = feat['hypertension'].apply(lambda x: round(x))
feat['heart_disease'] = feat['heart_disease'].apply(lambda x: round(x))
feat['ever_married'] = feat['ever_married'].apply(lambda x: round(x))
# 分割训练和测试集
X_train, X_test, y_train, y_test = train_test_split(feat, target, test_size=0.3, random_state=2)

这是问题的一部分。

Kfold = KFold(n_splits=10)
pipeline = make_pipeline(StandardScaler(), DecisionTreeClassifier())
n_iter = 0
for train_idx, test_idx in Kfold.split(feat):
    pipeline.fit(X_train[train_idx], y_train[train_idx])
    score = pipeline.score(X_train[test_idx], y_train[test_idx])
    print('Fold #{} accuracy{}'.format(1, score))

错误代码

Traceback (most recent call last):
  File "/Users/merb/Documents/Dev/DataScience/TP.py", line 84, in <module>
    pipeline.fit(X_train[train_idx], y_train[train_idx])
  File "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/site-packages/pandas/core/frame.py", line 3030, in __getitem__
    indexer = self.loc._get_listlike_indexer(key, axis=1, raise_missing=True)[1]
  File "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/site-packages/pandas/core/indexing.py", line 1266, in _get_listlike_indexer
    self._validate_read_indexer(keyarr, indexer, axis, raise_missing=raise_missing)
  File "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/site-packages/pandas/core/indexing.py", line 1308, in _validate_read_indexer
    raise KeyError(f"None of [{key}] are in the [{axis_name}]")
KeyError: "None of [Int64Index([ 5893,  5894,  5895,  5896,  5897,  5898,  5899,  5900,    5901,\n             5902,\n            ...\n            58912, 58913, 58914, 58915, 58916, 58917, 58918, 58919, 58920,\n            58921],\n           dtype='int64', length=53029)] are in the [columns]"

回答：

您应该使用df.loc[indexes]来按索引选择行。如果您想按整数位置选择行，您应该使用df.iloc[indexes]。

此外，您可以阅读这个页面来了解如何使用pandas进行索引和选择数据。