From 7193c1354b75f3e74a918fd7271e2f69ccf8b7fc Mon Sep 17 00:00:00 2001 From: DREMN2112 Date: Thu, 7 Nov 2024 17:09:10 +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 | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/function.py b/function.py index b4dc1d9..10f575b 100644 --- a/function.py +++ b/function.py @@ -10,15 +10,16 @@ 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, lambda_reg=0.1): """ - 对率回归训练函数。 + 对率回归训练函数，加入了 L2 正则化。 Args: - X (numpy array): 特征数据。 - y (numpy array): 标签数据。 - learning_rate (float): 学习率，默认值为 0.01。 - iterations (int): 梯度下降迭代次数，默认值为 1000。 + - lambda_reg (float): 正则化强度，默认值为 0.1。 Returns: - numpy array: 训练得到的权重。 @@ -34,13 +35,13 @@ def train_logreg(X, y, learning_rate=0.01, iterations=1000): # 计算误差 error = predictions - y - # 更新权重 - gradient = np.dot(X.T, error) / len(y) + # 计算梯度，并加入正则化项 + gradient = np.dot(X.T, error) / len(y) + lambda_reg * weights / len(y) weights -= learning_rate * gradient # 每 100 次迭代输出一次损失 if i % 100 == 0: - loss = -np.mean(y * np.log(predictions) + (1 - y) * np.log(1 - predictions)) + loss = -np.mean(y * np.log(predictions) + (1 - y) * np.log(1 - predictions)) + (lambda_reg / (2 * len(y))) * np.sum(weights**2) print(f"Iteration {i}, Loss: {loss:.4f}") return weights @@ -101,8 +102,8 @@ if __name__ == "__main__": # 加载测试集 X_test, y_test = load_data("./data/horse-colic.test") - # 训练对率回归模型，增加迭代次数为 5000 次 - weights = train_logreg(X_train, y_train, learning_rate=0.01, iterations=5000) + # 训练对率回归模型，加入正则化并增加迭代次数为 5000 次 + weights = train_logreg(X_train, y_train, learning_rate=0.01, iterations=5000, lambda_reg=0.1) # 测试单个数据 y_pred = predict_logreg(X_test[11], weights)