Google Colab泰坦尼克号ML项目报错：无法将字节/字符串转数值类型

错误原因 #

GaussianNB（高斯朴素贝叶斯）模型仅支持数值型输入特征，你的训练集特征label_train中存在字符串/字节类型的列，模型无法自动将这类非数值数据转换为可计算的数值，因此触发该错误。

解决步骤 #

1. 定位非数值列

先检查哪些特征是字符串类型：

print(label_train.dtypes)

输出结果中，dtype为object的列就是需要处理的非数值列。

2. 转换非数值特征

根据特征类型选择合适的编码方式：

方式一：独热编码（适合无序类别特征，如性别、舱位等级）

用pandas.get_dummies自动处理所有类别型特征：

# 替换原特征集为编码后的版本
label_train = pd.get_dummies(label_train)

# 后续训练代码保持不变
x_train, x_test, y_train, y_test = train_test_split(label_train, feature_train, test_size=0.2)
clf.fit(x_train, np.ravel(y_train))

方式二：标签编码（适合有序类别特征，如学历等级、会员等级）

用sklearn.preprocessing.LabelEncoder对每一列字符串特征编码：

from sklearn.preprocessing import LabelEncoder

le = LabelEncoder()
for col in label_train.columns:
    if label_train[col].dtype == 'object':
        # 注意：若列存在空字符串，先转为NaN再处理
        label_train[col] = label_train[col].replace('', np.nan)
        label_train[col] = le.fit_transform(label_train[col].astype(str))

3. 额外检查：处理隐性缺失值

部分数据中可能存在空字符串（而非NaN），这类值也会被识别为字符串类型，需要先转换为NaN再处理：

# 将所有空字符串替换为NaN
label_train = label_train.replace('', np.nan)
# 再删除含缺失值的行（或用fillna填充）
label_train = label_train.dropna()

修正后的完整代码示例 #

import pandas as pd
import numpy as np
from sklearn.naive_bayes import GaussianNB
from sklearn.model_selection import train_test_split, cross_val_score, cross_val_predict
from sklearn.metrics import confusion_matrix
import seaborn as sns
import matplotlib.pyplot as plt

# Data to be predicted
X_to_be_predicted = my_df[my_df.Survived.isnull()]
X_to_be_predicted = X_to_be_predicted.drop(['Survived'], axis = 1)
# X_to_be_predicted[X_to_be_predicted.Age.isnull()]
# X_to_be_predicted.dropna(inplace = True) # 417 x 27

#Training data
train_data = my_df
train_data = train_data.dropna()
feature_train = train_data['Survived']
label_train = train_data.drop(['Survived'], axis = 1)

# --- 新增特征编码步骤 ---
# 用独热编码处理类别特征
label_train = pd.get_dummies(label_train)

##Gaussian
clf = GaussianNB()
x_train, x_test, y_train, y_test = train_test_split(label_train, feature_train, test_size=0.2)
clf.fit(x_train, np.ravel(y_train))
print("NB Accuracy: " + repr(round(clf.score(x_test, y_test) * 100, 2)) + "%")
result_rf = cross_val_score(clf, x_train, y_train, cv=10, scoring = 'accuracy')
print('The cross validated score for GaussianNB is:', round(result_rf.mean()*100,2))
y_pred = cross_val_predict(clf,x_train, y_train, cv=10)
sns.heatmap(confusion_matrix(y_train, y_pred), annot=True, fmt = '3.0f', cmap = "summer")
plt.title('Confusion_matrix for NB', y=1.05, size=15)
plt.show()

内容的提问来源于stack exchange，提问作者Shyngys Nurgaliev