From 8f57fdf0cc5639ae78f7066642b1d475449ca00d Mon Sep 17 00:00:00 2001
From: YimingTang <Clocky7@outlook.com>
Date: Thu, 24 Oct 2024 17:28:36 +0800
Subject: [PATCH] =?UTF-8?q?=E6=9B=B4=E6=96=B0=20function.py?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 function.py | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)

diff --git a/function.py b/function.py
index 409b1ac..b176385 100644
--- a/function.py
+++ b/function.py
@@ -21,8 +21,8 @@ import numpy as np
 """
 请在这里填写学号和姓名，不填写即成绩无效。
 """
-id = "202210423xxx"
-name = "test_name"
+id = "202210423218"
+name = "唐一鸣"
 
 
 # 这里写自己的代码实现
@@ -38,6 +38,17 @@ def train_logreg(X, y, other_params="你自己设置一些其它参数"):
     Returns:
     - numpy array: 训练得到的权重。
     """
+    m, n = X.shape  # 样本数和特征数
+    weights = np.zeros(n)  # 初始化权重
+    for _ in range(iterations):
+        # 计算当前模型的预测值
+        z = np.dot(X, weights)
+        h = sigmoid(z)
+        # 计算损失函数的梯度
+        gradient = np.dot(X.T, (h - y)) / m
+        # 更新权重
+        weights -= learning_rate * gradient
+    return weights
 
 
 def predict_logreg(inX, weights):
@@ -51,7 +62,8 @@ def predict_logreg(inX, weights):
     Returns:
     - int: 预测的分类标签。
     """
-
+    probability = sigmoid(np.dot(inX, weights))
+    return 1 if probability >= 0.5 else 0
 
 # 这里也写自己的代码实现
 def load_data(filepath):
@@ -65,6 +77,10 @@ def load_data(filepath):
     - np.array: 特征数据。
     - np.array: 标签数据。
     """
+    data = np.loadtxt(filepath, delimiter=' ', dtype=float)
+    X = data[:, :-1]  # 特征数据
+    y = data[:, -1]   # 标签数据
+    return X, y
 
 
 if __name__ == "__main__":