generated from LufsX/Experiment-LogReg
This commit is contained in:
+21
-39
@@ -18,22 +18,17 @@ import pandas as pd
|
|||||||
- numpy: 2.1.1
|
- numpy: 2.1.1
|
||||||
"""
|
"""
|
||||||
|
|
||||||
|
|
||||||
"""
|
"""
|
||||||
请在这里填写学号和姓名,不填写即成绩无效。
|
请在这里填写学号和姓名,不填写即成绩无效。
|
||||||
姓名请直接**填写中文**
|
姓名请直接**填写中文**
|
||||||
例如: name = "张三"
|
例如: name = "张三"
|
||||||
"""
|
"""
|
||||||
id = "202210423223"
|
id = "20221042323"
|
||||||
name = "李顺潮"
|
name = "李顺潮"
|
||||||
|
|
||||||
def sigmoid(z):
|
|
||||||
"""Sigmoid函数,用于对率回归的预测。"""
|
|
||||||
return 1 / (1 + np.exp(-z))
|
|
||||||
|
|
||||||
|
|
||||||
# 这里写自己的代码实现
|
# 这里写自己的代码实现
|
||||||
def train_logreg(X, y, other_params="你自己设置一些其它参数"):
|
def train_logreg(X, y, other_params=None):
|
||||||
"""
|
"""
|
||||||
对率回归训练函数。
|
对率回归训练函数。
|
||||||
|
|
||||||
@@ -45,17 +40,17 @@ def train_logreg(X, y, other_params="你自己设置一些其它参数"):
|
|||||||
Returns:
|
Returns:
|
||||||
- numpy array: 训练得到的权重。
|
- numpy array: 训练得到的权重。
|
||||||
"""
|
"""
|
||||||
# 初始化权重
|
m, n = X.shape
|
||||||
weights = np.zeros(X.shape[1])
|
weights = np.zeros(n) # 初始化权重
|
||||||
learning_rate = 0.01
|
learning_rate = 0.01 # 学习率设置
|
||||||
iterations = 1000
|
num_iterations = 1000 # 迭代次数
|
||||||
|
|
||||||
for _ in range(iterations):
|
# 对率回归的梯度下降
|
||||||
z = np.dot(X, weights)
|
for _ in range(num_iterations):
|
||||||
predictions = sigmoid(z)
|
linear_model = np.dot(X, weights)
|
||||||
error = y - predictions
|
predictions = 1 / (1 + np.exp(-linear_model)) # Sigmoid函数
|
||||||
gradient = np.dot(X.T, error)
|
errors = y - predictions
|
||||||
weights += learning_rate * gradient
|
weights += learning_rate * np.dot(X.T, errors) / m # 更新权重
|
||||||
|
|
||||||
return weights
|
return weights
|
||||||
|
|
||||||
@@ -71,10 +66,9 @@ def predict_logreg(inX, weights):
|
|||||||
Returns:
|
Returns:
|
||||||
- int: 预测的分类标签。
|
- int: 预测的分类标签。
|
||||||
"""
|
"""
|
||||||
# 计算预测值
|
linear_model = np.dot(inX, weights)
|
||||||
z = np.dot(inX, weights)
|
predictions = 1 / (1 + np.exp(-linear_model)) # Sigmoid函数
|
||||||
probability = sigmoid(z)
|
return np.round(predictions).astype(int) # 返回0或1
|
||||||
return 1 if probability >= 0.5 else 0
|
|
||||||
|
|
||||||
|
|
||||||
# 这里也写自己的代码实现
|
# 这里也写自己的代码实现
|
||||||
@@ -89,17 +83,13 @@ def load_data(filepath):
|
|||||||
- np.array: 特征数据。
|
- np.array: 特征数据。
|
||||||
- np.array: 标签数据。
|
- np.array: 标签数据。
|
||||||
"""
|
"""
|
||||||
# 使用pandas加载数据
|
data = pd.read_csv(filepath, header=None)
|
||||||
df = pd.read_csv(filepath, header=None)
|
# 假设最后一列是标签,其余为特征
|
||||||
|
X = data.iloc[:, :-1].values
|
||||||
# 替换缺失值和非数值数据,这里假设我们用0来填补缺失数据
|
y = data.iloc[:, -1].values
|
||||||
df.replace('?', np.nan, inplace=True)
|
|
||||||
df.fillna(0, inplace=True)
|
|
||||||
|
|
||||||
# 将数据拆分为特征和标签
|
|
||||||
X = df.iloc[:, :-1].values # 特征
|
|
||||||
y = df.iloc[:, -1].values # 标签
|
|
||||||
|
|
||||||
|
# 转换标签为数值,如果是字符串类型可以这样处理
|
||||||
|
y = np.where(y == '1', 1, 0) # 假设1表示死亡,0表示存活
|
||||||
return X, y
|
return X, y
|
||||||
|
|
||||||
|
|
||||||
@@ -107,11 +97,3 @@ if __name__ == "__main__":
|
|||||||
# 以下代码可用于测试,可自行随意修改
|
# 以下代码可用于测试,可自行随意修改
|
||||||
# 加载训练集
|
# 加载训练集
|
||||||
X_train, y_train = load_data("./data/horse-colic.data")
|
X_train, y_train = load_data("./data/horse-colic.data")
|
||||||
|
|
||||||
# 训练模型
|
|
||||||
weights = train_logreg(X_train, y_train)
|
|
||||||
|
|
||||||
# 测试单个样本的预测
|
|
||||||
sample = X_train[0].reshape(1, -1) # 取出第一行作为样本
|
|
||||||
prediction = predict_logreg(sample, weights)
|
|
||||||
print(f"样本预测: {prediction}")
|
|
||||||
|
|||||||
Reference in New Issue
Block a user