python写softmax分类三类平面上的二维点的代码

好的，以下是Python写softmax分类三类平面上的二维点的代码：

import numpy as np

# 随机生成三类平面上的二维点
num_points = 100
dimensions = 2
points = np.random.randint(0, 100, [num_points, dimensions])
labels = np.random.randint(0, 3, [num_points, 1])

# 将标签转化为独热编码
one_hot_labels = np.zeros((num_points, 3))
for i in range(num_points):
    one_hot_labels[i, labels[i]] = 1

# 初始化权重矩阵和偏置向量
weights = np.random.randn(dimensions, 3)
biases = np.zeros((1, 3))

# 定义softmax函数
def softmax(x):
    return np.exp(x) / np.sum(np.exp(x), axis=0)

# 定义交叉熵损失函数
def cross_entropy_loss(predicted, actual):
    loss = -np.sum(actual * np.log(predicted))
    return loss

# 定义学习率和训练次数
learning_rate = 0.01
epochs = 1000

# 训练模型
for epoch in range(epochs):
    # 前向传播
    predicted = np.dot(points, weights) + biases
    predicted = softmax(predicted)

    # 计算损失
    loss = cross_entropy_loss(predicted, one_hot_labels)

    # 反向传播
    error = predicted - one_hot_labels
    dw = np.dot(points.T, error)
    db = np.sum(error, axis=0, keepdims=True)

    # 更新权重和偏置
    weights -= learning_rate * dw
    biases -= learning_rate * db

    # 每100次迭代输出一次损失
    if epoch % 100 == 0:
        print('Epoch %d loss: %.4f' % (epoch, loss))

希望对您有帮助！