### 基于 Softmax 回归实现鸢尾花数据集分类的 Jupyter Notebook 代码示例 以下是一个完整的基于 TensorFlow 和 Keras 实现 Softmax 回归对鸢尾花数据集进行分类的代码示例。该代码涵盖了数据加载、预处理、模型构建、编译、训练和评估的过程。 ```python ### 导入必要的库 import tensorflow as tf from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Dense from sklearn.datasets import load_iris from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler, OneHotEncoder import numpy as np ### 加载鸢尾花数据集 data = load_iris() X = data.data # 特征 y = data.target # 标签 ### 将标签转换为独热编码形式 one_hot_encoder = OneHotEncoder(sparse=False) y_onehot = one_hot_encoder.fit_transform(y.reshape(-1, 1)) ### 划分训练集和测试集 X_train, X_test, y_train, y_test = train_test_split(X, y_onehot, test_size=0.2, random_state=42) ### 对特征进行标准化处理 scaler = StandardScaler() X_train_scaled = scaler.fit_transform(X_train) X_test_scaled = scaler.transform(X_test) ### 构建 Softmax 回归模型 model = Sequential([ Dense(3, activation='softmax', input_shape=(X_train.shape[1],)) # 输出层有 3 个神经元对应三个类别 ]) ### 编译模型 model.compile( optimizer='adam', loss='categorical_crossentropy', # 使用交叉熵损失函数 metrics=['accuracy'] ) ### 打印模型结构 model.summary() ### 训练模型 history = model.fit( X_train_scaled, y_train, epochs=100, batch_size=16, validation_data=(X_test_scaled, y_test), verbose=2 ) ### 测试模型性能 loss, accuracy = model.evaluate(X_test_scaled, y_test, verbose=2) print(f'Test Accuracy: {accuracy:.4f}') # 显示测试集上的准确率[^6] ### 进行预测 predictions = model.predict(X_test_scaled[:5]) # 预测前五个样本 predicted_classes = np.argmax(predictions, axis=1) # 获取最大概率对应的类别索引 true_classes = np.argmax(y_test[:5], axis=1) # 获取真实类别 for i in range(len(predicted_classes)): print(f'Predicted Class: {predicted_classes[i]}, True Class: {true_classes[i]}') ``` --- ###