机器学习_Logistic回归简单交叉验证

# ================ 关于泰坦尼克号乘客的生还率 ================
# 加载包
import pandas as pd
from sklearn.linear_model import LogisticRegression  # Logistic 回归模型 包
from sklearn.linear_model import LogisticRegressionCV # 带有正则化参数C的粒度
from sklearn.model_selection import cross_val_score # 交叉验证

# 文件位置
fil_tr = r"E:\python Data\happytry\Kaggle_Titanic-master\train.csv"
fil_te = r"E:\python Data\happytry\Kaggle_Titanic-master\test.csv"

# 加载数据
train = pd.read_csv(fil_tr)
test = pd.read_csv(fil_te)

# ====================================pandas预处理=================================
# 查看数据
train[:4] 
## statsmodels 和 scikit-learn通常不能接收缺失数据
train.isnull().sum()
test.isnull().sum()

# 用年龄预测生存率
## 补全缺失
train['Age'] = train['Age'].fillna(train['Age'].median())
test['Age'] = test['Age'].fillna(test['Age'].median())
## 将性别转换为 0 1变量
train['IsFemale'] = (train['Sex'] == 'female') * 1
test['IsFemale'] = (test['Sex'] == 'female') * 1
# =================================================================================

# ====================================转换成numpy进入模型============================
predictors = ['Pclass', 'IsFemale', 'Age']
## 装换为numpy array
x_train = train[predictors].values
x_test = test[predictors].values
y_train = train['Survived'].values

# ===== 创建Logistic模型 ======
model = LogisticRegression()
model.fit(x_train, y_train)
# 预测 
## 训练集预测
y_predict = model.predict(x_train)
## 测试集预测
y_predict1 = model.predict(x_test)
# 计算错误率
(train['Survived'] == y_predict).mean()

# logisticregressioncv 类用参数指定网格搜索对模型的正则化参数C的粒度
model_cv = LogisticRegressionCV(10)
model_cv.fit(x_train, y_train)

# 交叉验证带有四个不重叠训练数据的模型
model = LogisticRegression(C = 10)
scores = cross_val_score(model, x_train, y_train, cv = 4) # 做四次交叉验证