From b880be4230a183413bfa8d3a225f00461ec9553c Mon Sep 17 00:00:00 2001
From: wzy <675163717@qq.com>
Date: Thu, 24 Oct 2024 17:22:19 +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 | 43 +++++++++++++++++++++++++++++++++++--------
 1 file changed, 35 insertions(+), 8 deletions(-)

diff --git a/function.py b/function.py
index da19929..b5f0967 100644
--- a/function.py
+++ b/function.py
@@ -29,16 +29,29 @@ name = "万子玉"
 def sigmoid(z):
     return 1 / (1 + np.exp(-z))
 
-def train_logreg(X, y, learning_rate=0.01, iterations=1000):
+def train_logreg(X, y, learning_rate=0.01, iterations=1000, batch_size=32, initial_weights=None):
     X = np.insert(X, 0, 1, axis=1)  # 添加偏置项
-    weights = np.zeros(X.shape[1])
+    if initial_weights is None:
+        weights = np.zeros(X.shape[1])
+    else:
+        weights = initial_weights
+    
     m = X.shape[0]
     
     for _ in range(iterations):
-        z = np.dot(X, weights)
-        h = sigmoid(z)
-        gradient = np.dot(X.T, (h - y)) / m
-        weights -= learning_rate * gradient
+        indices = np.arange(m)
+        np.random.shuffle(indices)
+        X_shuffled = X[indices]
+        y_shuffled = y[indices]
+        
+        for i in range(0, m, batch_size):
+            end = i + batch_size
+            X_batch = X_shuffled[i:end]
+            y_batch = y_shuffled[i:end]
+            z = np.dot(X_batch, weights)
+            h = sigmoid(z)
+            gradient = np.dot(X_batch.T, (h - y_batch)) / batch_size
+            weights -= learning_rate * gradient
         
     return weights
     """
@@ -105,15 +118,29 @@ def load_data(filepath):
 
 
 if __name__ == "__main__":
-    # 以下代码可用于测试，可自行随意修改
     # 加载训练集
     X_train, y_train = load_data("./data/horse-colic.data")
 
     # 加载测试集
     X_test, y_test = load_data("./data/horse-colic.test")
 
+    # 模型调优参数
+    learning_rates = [0.01, 0.001, 0.0001]
+    iterations = [500, 1000, 2000]
+    batch_sizes = [16, 32, 64]
+    initial_weights = np.random.randn(X_train.shape[1] + 1)  # 随机初始化权重
+
+    # 选择一组参数进行训练
+    chosen_learning_rate = learning_rates[1]
+    chosen_iterations = iterations[1]
+    chosen_batch_size = batch_sizes[1]
+
     # 训练对率回归模型
-    weights = train_logreg(X_train, y_train, learning_rate=0.01, iterations=1000)
+    weights = train_logreg(X_train, y_train, 
+                           learning_rate=chosen_learning_rate, 
+                           iterations=chosen_iterations, 
+                           batch_size=chosen_batch_size, 
+                           initial_weights=initial_weights)
 
     # 测试单个数据
     y_pred = predict_logreg(X_test[11], weights)