执行独热编码后仍出现could not convert string to float错误求助

你的错误根源是编码后没有替换原数据中的字符串特征列，反而把编码后的列和原字符串列拼接在一起，导致最终用于训练的X_train仍然包含gender这类字符串型特征（比如'Male'），而RandomForestClassifier只能处理数值型输入。

另外你混淆了LabelEncoder的用法：LabelEncoder通常用于**目标变量（y）**的编码，而非特征；如果是特征编码，对于无序分类特征（如gender、work_type），更适合用独热编码（OneHotEncoder）或pd.get_dummies，避免引入不必要的顺序关系。

方案一：用pd.get_dummies快速处理独热编码 #

直接替换原有的编码逻辑，用pd.get_dummies自动识别分类列并编码，同时保留数值列：

import pandas as pd
import numpy as np
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score

raw_data = pd.read_csv('/Users/name/Desktop/CompSci_Projects/Stroke_Prediction_Model/dataset/healthcare-dataset-stroke-data.csv')
data = raw_data.copy(deep=True)
data.interpolate(method='linear', inplace=True)

# 处理分类特征：用pd.get_dummies做独热编码，排除id和目标列
categorical_cols = ['gender', 'ever_married', 'work_type', 'Residence_type']
data_encoded = pd.get_dummies(data, columns=categorical_cols, drop_first=True)

# 拆分已知/未知smoking_status的数据
smokestatus_known = data_encoded[data_encoded['smoking_status'] != 'Unknown']
smokestatus_unknown = data_encoded[data_encoded['smoking_status'] == 'Unknown']

# 定义特征和目标
x_columns_to_drop = ['id', 'smoking_status']
y_column = 'smoking_status'

X_train = smokestatus_known.drop(x_columns_to_drop, axis=1)
y_train = smokestatus_known[y_column]

X_test = smokestatus_unknown.drop(x_columns_to_drop, axis=1)

# 训练模型
classifier = RandomForestClassifier()
classifier.fit(X_train, y_train)

# 预测未知值
predicted_values = classifier.predict(X_test)

方案二：用sklearn.ColumnTransformer规范编码（更适合机器学习流水线） #

如果需要更规范的流程（比如后续要做交叉验证），推荐用ColumnTransformer来分别处理数值和分类特征：

import pandas as pd
import numpy as np
from sklearn.ensemble import RandomForestClassifier
from sklearn.preprocessing import OneHotEncoder
from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline

raw_data = pd.read_csv('/Users/name/Desktop/CompSci_Projects/Stroke_Prediction_Model/dataset/healthcare-dataset-stroke-data.csv')
data = raw_data.copy(deep=True)
data.interpolate(method='linear', inplace=True)

# 拆分已知/未知smoking_status的数据
smokestatus_known = data[data['smoking_status'] != 'Unknown']
smokestatus_unknown = data[data['smoking_status'] == 'Unknown']

# 定义特征列类型
categorical_cols = ['gender', 'ever_married', 'work_type', 'Residence_type']
numeric_cols = ['age', 'hypertension', 'heart_disease', 'avg_glucose_level', 'bmi', 'stroke']

# 构建预处理流水线：分类列独热编码，数值列保持不变
preprocessor = ColumnTransformer(
    transformers=[
        ('cat', OneHotEncoder(drop='first'), categorical_cols),
        ('num', 'passthrough', numeric_cols)
    ])

# 定义模型流水线
model = Pipeline(steps=[
    ('preprocessor', preprocessor),
    ('classifier', RandomForestClassifier())
])

# 准备训练数据
X_train = smokestatus_known.drop(['id', 'smoking_status'], axis=1)
y_train = smokestatus_known['smoking_status']

X_test = smokestatus_unknown.drop(['id', 'smoking_status'], axis=1)

# 训练和预测
model.fit(X_train, y_train)
predicted_values = model.predict(X_test)

额外注意点 #

1. 你原代码中train_test_split的结果被后续的X_train = x_train覆盖了，这会导致你没有做验证集拆分，无法评估模型性能。如果需要评估，建议保留验证集逻辑。
2. smoking_status作为目标变量，sklearn的分类器支持直接处理字符串标签，不需要额外编码。

内容的提问来源于stack exchange，提问作者rts2027