河南大学-机器学习-监督学习综合实验

实验目的

1. 掌握数据预处理方法；

      2.通过参数寻优，构造最优分类模型。

实验内容

1. 基于泰坦尼克号数据集，构建一个可以根据乘客个人信息，如性别、年龄、船舱等级等来推测乘客是否生存的分类模型。要求：

(1)在学习过的分类算法中至少选择3种分别进行模型的选择、训练和预测；

(2)使用Scikit-learn提供的GridSearchCV函数进行参数寻优；

(3)test.csv文件中没有标签列，可以只使用训练数据集文件train.csv进行

训练集和测试集的划分、训练、测试与评价；

(4)将数据集test.csv的预测结果上传到kaggle平台进行检验的请在WORD文

档中说明并截图，会有额外加分。

3.代码

import pandas as pd
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score
from sklearn.impute import SimpleImputer
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder

# 加载数据集
train_data = pd.read_csv(r'F:\File\大三下\机器学习\监督学习综合实验\titanic_train.csv')
test_data = pd.read_csv(r'F:\File\大三下\机器学习\监督学习综合实验\titanic_test.csv')


# 合并训练集和测试集以便进行统一的预处理
full_data = pd.concat([train_data.drop('Survived', axis=1), test_data], axis=0)

# 对数据进行预处理和特征选择
X = full_data.drop(['PassengerId', 'Name', 'Ticket', 'Cabin'], axis=1)
y_train = train_data['Survived']

# 将类别型特征转换为数值型特征
categorical_features = ['Sex', 'Embarked']
numeric_features = ['Age', 'Fare', 'Pclass', 'SibSp', 'Parch']

# 处理缺失值
numeric_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='median')),
    ('scaler', StandardScaler())])

categorical_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='most_frequent')),
    ('onehot', OneHotEncoder(handle_unknown='ignore'))])

preprocessor = ColumnTransformer(
    transformers=[
        ('num', numeric_transformer, numeric_features),
        ('cat', categorical_transformer, categorical_features)])

# 对整个数据集进行拟合和转换
X_processed = preprocessor.fit_transform(X)

# 划分训练集和验证集
X_train = X_processed[:len(train_data)]
X_test = X_processed[len(train_data):]

X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.2, random_state=42)

# 定义分类器
classifiers = {
    'Random Forest': RandomForestClassifier(),
    'SVM': SVC(),
    'Logistic Regression': LogisticRegression()
}

# 定义参数网格
param_grids = {
    'Random Forest': {'classifier__n_estimators': [100, 200, 300], 'classifier__max_depth': [5, 10, 15]},
    'SVM': {'classifier__C': [0.1, 1, 10], 'classifier__gamma': [0.1, 0.01, 0.001], 'classifier__kernel': ['rbf']},
    'Logistic Regression': {'classifier__C': [0.1, 1, 10], 'classifier__penalty': ['l2']}
}

# 创建Pipeline并使用GridSearchCV进行参数搜索和模型选择
results = {}
for clf_name, clf in classifiers.items():
    pipe = Pipeline(steps=[('classifier', clf)])
    param_grid = param_grids[clf_name]
    grid_search = GridSearchCV(pipe, param_grid, cv=5)
    grid_search.fit(X_train, y_train)
    results[clf_name] = grid_search

# 输出最优模型的结果
for clf_name, result in results.items():
    print(f"{clf_name}: Best parameters - {result.best_params_}, Best score - {result.best_score_}")

# 使用最优模型在验证集上进行预测和评价
best_clf = results['Random Forest'].best_estimator_
y_pred = best_clf.predict(X_val)
accuracy = accuracy_score(y_val, y_pred)
print(f"Validation Accuracy: {accuracy}")

# 在测试集上进行预测
y_test_pred = best_clf.predict(X_test)

# 生成提交文件
submission = pd.DataFrame({'PassengerId': test_data['PassengerId'], 'Survived': y_test_pred})
submission.to_csv('submission.csv', index=False)