Python打卡：Day24

import pandas as pd
import numpy as np
import re
import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix, accuracy_score, precision_score, recall_score, f1_score
import warnings
 
warnings.filterwarnings("ignore")
 
# 统一的预处理函数
def preprocess_data(df):
    # 1. 填充Age缺失值（用训练集中位数填充）
    median_age = preprocess_data.median_age if hasattr(preprocess_data, 'median_age') else df['Age'].median()
    df['Age'].fillna(median_age, inplace=True)
    preprocess_data.median_age = median_age  # 记录中位数以便测试集使用
 
    # 2. 填充Embarked缺失值（用训练集众数填充）
    mode_embarked = preprocess_data.mode_embarked if hasattr(preprocess_data, 'mode_embarked') else df['Embarked'].mode()[0]
    df['Embarked'].fillna(mode_embarked, inplace=True)
    preprocess_data.mode_embarked = mode_embarked
 
    # 3. 新增Deck列（取Cabin首字母），缺失填'Unknown'
    df['Deck'] = df['Cabin'].str[0].fillna('Unknown')
 
    # 4. 新增Cabin_Missing列（Cabin是否缺失）
    df['Cabin_Missing'] = df['Cabin'].isnull().astype(int)
 
    # 5. 删除Cabin列
    df.drop(columns=['Cabin'], inplace=True)
 
    # 6. 性别转换为数字列 is_male
    df['is_male'] = (df['Sex'] == 'male').astype(int)
    df.drop(columns=['Sex'], inplace=True)
 
    # 7. Embarked独热编码
    embarked_dummies = pd.get_dummies(df['Embarked'], prefix='Embarked', dtype=int)
    df = pd.concat([df, embarked_dummies], axis=1)
    df.drop(columns=['Embarked'], inplace=True)
 
    # 8. Ticket类型特征提取
    def ticket_type(ticket):
        if pd.isnull(ticket):
            return 'Unknown'
        elif ticket.isdigit():
            return 'Numeric'
        elif re.match(r'^[A-Za-z\. ]+', ticket):
            return 'Alphanumeric'
        else:
            return 'Other'
    df['TicketType'] = df['Ticket'].apply(ticket_type)
    ticket_counts = df['Ticket'].value_counts()
    df['sharing_number'] = df['Ticket'].map(ticket_counts)
    df.drop(columns=['Ticket'], inplace=True)
    ticket_dummies = pd.get_dummies(df['TicketType'], prefix='Ticket', dtype=int)
    df = pd.concat([df, ticket_dummies], axis=1)
    df.drop(columns=['TicketType'], inplace=True)
 
    # 9. Deck映射为序号，缺失用众数填充
    deck_order = {'A': 1, 'B': 2, 'C': 3, 'D': 4,
                  'E': 5, 'F': 6, 'G': 7, 'T': 8, 'Unknown': 0}
    df['Deck_Ordinal'] = df['Deck'].map(deck_order)
    mode_deck = preprocess_data.mode_deck if hasattr(preprocess_data, 'mode_deck') else df['Deck_Ordinal'].mode()[0]
    df['Deck_Ordinal'].fillna(mode_deck, inplace=True)
    preprocess_data.mode_deck = mode_deck
    df.drop(columns=['Deck'], inplace=True)
 
    # 10. 删除Name列（无用）
    if 'Name' in df.columns:
        df.drop(columns=['Name'], inplace=True)
 
    # 11. 删除PassengerId列（如果有，预测时用）
    # 这里不删，保留传外面
 
    return df
 
# === 读取训练数据 ===
train_data = pd.read_csv('train.csv')
 
# 记录训练集PassengerId（一般训练时不用）
if 'PassengerId' in train_data.columns:
    train_passenger_ids = train_data['PassengerId']
 
# 预处理训练数据
train_processed = preprocess_data(train_data)
 
# 准备训练特征和标签
X = train_processed.drop(columns=['Survived', 'PassengerId'], errors='ignore')
y = train_processed['Survived']
 
# 划分训练/验证集
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)
 
# === 训练XGBoost模型 ===
xgb_model = xgb.XGBClassifier(random_state=42, use_label_encoder=False, eval_metric='logloss')
xgb_model.fit(X_train, y_train)
 
# 验证集预测和评估
y_val_pred = xgb_model.predict(X_val)
print("验证集分类报告：")
print(classification_report(y_val, y_val_pred))
print("验证集混淆矩阵：")
print(confusion_matrix(y_val, y_val_pred))
 
# === 读取测试数据 ===
test_data = pd.read_csv('test.csv')
passenger_ids = test_data['PassengerId']  # 保留乘客ID用于提交
 
# 预处理测试数据（用训练集统计量填充）
test_processed = preprocess_data(test_data)
 
# 确保预测数据列与训练特征一致，删除PassengerId列
test_processed = test_processed.drop(columns=['PassengerId'], errors='ignore')
 
# 对测试数据的列补齐（训练集里有，测试集没的列补0）
for col in X.columns:
    if col not in test_processed.columns:
        test_processed[col] = 0
 
# 测试数据列顺序对齐训练集
test_processed = test_processed[X.columns]
 
# 预测测试集
test_preds = xgb_model.predict(test_processed)
 
# 保存预测结果到csv
submission = pd.DataFrame({
    'PassengerId': passenger_ids,
    'Survived': test_preds
})
submission.to_csv('submission.csv', index=False)
print("预测结果已保存到 submission.csv")

@浙大疏锦行