机器学习：网格搜索参数优化

TensorFlow/PyTorch 进阶实现

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
import torch
import torch.nn as nn
import torch.optim as optim

# TensorFlow 实现
class TensorFlowModel:
    def __init__(self, input_dim: int, hidden_units: List[int] = [64, 32]):
        self.model = self._build_model(input_dim, hidden_units)

    def _build_model(self, input_dim: int, hidden_units: List[int]) -> keras.Model:
        inputs = keras.Input(shape=(input_dim,))
        x = inputs
        for units in hidden_units:
            x = layers.Dense(units, activation='relu')(x)
            x = layers.BatchNormalization()(x)
            x = layers.Dropout(0.2)(x)
        outputs = layers.Dense(1)(x)
        model = keras.Model(inputs=inputs, outputs=outputs)
        model.compile(optimizer=keras.optimizers.Adam(learning_rate=0.001), loss='mse', metrics=['mae'])
        return model

    def train(self, X_train, y_train, X_val, y_val, epochs=100, batch_size=32):
        history = self.model.fit(X_train, y_train, validation_data=(X_val, y_val), epochs=epochs, batch_size=batch_size, verbose=1)
        return history

    def predict(self, X):
        return self.model.predict(X)

# PyTorch 实现
class PyTorchModel(nn.Module):
    def __init__(self, input_dim: int, hidden_units: List[int] = [64, 32]):
        super(PyTorchModel, self).__init__()
        layers_list = []
        prev_units = input_dim
        for units in hidden_units:
            layers_list.append(nn.Linear(prev_units, units))
            layers_list.append(nn.ReLU())
            layers_list.append(nn.BatchNorm1d(units))
            layers_list.append(nn.Dropout(0.2))
            prev_units = units
        layers_list.append(nn.Linear(prev_units, 1))
        self.network = nn.Sequential(*layers_list)

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        return self.network(x)

    def train_model(self, train_loader, val_loader, epochs=100, lr=0.001):
        criterion = nn.MSELoss()
        optimizer = optim.Adam(self.parameters(), lr=lr)
        train_losses = []
        val_losses = []
        for epoch in range(epochs):
            self.train()
            train_loss = 0.0
            for X_batch, y_batch in train_loader:
                optimizer.zero_grad()
                outputs = self(X_batch)
                loss = criterion(outputs, y_batch)
                loss.backward()
                optimizer.step()
                train_loss += loss.item()
            self.eval()
            val_loss = 0.0
            with torch.no_grad():
                for X_batch, y_batch in val_loader:
                    outputs = self(X_batch)
                    loss = criterion(outputs, y_batch)
                    val_loss += loss.item()
            train_losses.append(train_loss / len(train_loader))
            val_losses.append(val_loss / len(val_loader))
            if (epoch + 1) % 10 == 0:
                print(f"Epoch {epoch+1}/{epochs}, Train Loss: {train_losses[-1]:.4f}, Val Loss: {val_losses[-1]:.4f}")
        return train_losses, val_losses

数据处理流程

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.impute import SimpleImputer
from typing import List, Tuple

class DataProcessor:
    def __init__(self):
        self.scaler = StandardScaler()
        self.label_encoders = {}
        self.imputer = SimpleImputer(strategy='mean')

    def process(self, data: pd.DataFrame, target_col: str, categorical_cols: List[str] = None, test_size: float = 0.2) -> Tuple:
        X = data.drop(columns=[target_col])
        y = data[target_col]
        # 处理缺失值
        X_num = X.select_dtypes(include=[np.number])
        X[X_num.columns] = self.imputer.fit_transform(X_num)
        # 编码类别特征
        if categorical_cols:
            for col in categorical_cols:
                if col in X.columns:
                    le = LabelEncoder()
                    X[col] = le.fit_transform(X[col].astype(str))
                    self.label_encoders[col] = le
        # 标准化
        X_scaled = self.scaler.fit_transform(X)
        # 划分数据集
        X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=test_size, random_state=42)
        return X_train, X_test, y_train, y_test

if __name__ == "__main__":
    data = pd.DataFrame({
        'feature1': np.random.randn(1000),
        'feature2': np.random.randn(1000),
        'feature3': np.random.choice(['A', 'B', 'C'], 1000),
        'target': np.random.randn(1000)
    })
    processor = DataProcessor()
    X_train, X_test, y_train, y_test = processor.process(data, target_col='target', categorical_cols=['feature3'])
    print(f"训练集形状：{X_train.shape}")
    print(f"测试集形状：{X_test.shape}")

模型评估方法

from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score, confusion_matrix, classification_report, mean_squared_error, mean_absolute_error, r2_score
import matplotlib.pyplot as plt
import seaborn as sns

class ModelEvaluator:
    @staticmethod
    def evaluate_classification(y_true, y_pred, y_prob=None):
        metrics = {
            'accuracy': accuracy_score(y_true, y_pred),
            'precision': precision_score(y_true, y_pred, average='weighted'),
            'recall': recall_score(y_true, y_pred, average='weighted'),
            'f1': f1_score(y_true, y_pred, average='weighted')
        }
        if y_prob is not None:
            metrics['roc_auc'] = roc_auc_score(y_true, y_prob, multi_class='ovr')
        return metrics

    @staticmethod
    def evaluate_regression(y_true, y_pred):
        return {
            'mse': mean_squared_error(y_true, y_pred),
            'rmse': np.sqrt(mean_squared_error(y_true, y_pred)),
            'mae': mean_absolute_error(y_true, y_pred),
            'r2': r2_score(y_true, y_pred)
        }

    @staticmethod
    def plot_confusion_matrix(y_true, y_pred, labels=None):
        cm = confusion_matrix(y_true, y_pred)
        plt.figure(figsize=(8, 6))
        sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', xticklabels=labels, yticklabels=labels)
        plt.title('混淆矩阵')
        plt.xlabel('预测值')
        plt.ylabel('真实值')
        plt.show()

    @staticmethod
    def plot_learning_curve(train_losses, val_losses):
        plt.figure(figsize=(10, 6))
        plt.plot(train_losses, label='训练损失')
        plt.plot(val_losses, label='验证损失')
        plt.xlabel('Epoch')
        plt.ylabel('Loss')
        plt.title('学习曲线')
        plt.legend()
        plt.grid(True)
        plt.show()

if __name__ == "__main__":
    y_true_cls = [0, 1, 0, 1, 0, 1, 0, 0, 1, 1]
    y_pred_cls = [0, 1, 0, 0, 0, 1, 1, 0, 1, 1]
    cls_metrics = ModelEvaluator.evaluate_classification(y_true_cls, y_pred_cls)
    print("分类指标:", cls_metrics)
    y_true_reg = np.array([1.0, 2.0, 3.0, 4.0, 5.0])
    y_pred_reg = np.array([1.1, 1.9, 3.2, 3.8, 5.1])
    reg_metrics = ModelEvaluator.evaluate_regression(y_true_reg, y_pred_reg)
    print("回归指标:", reg_metrics)

实践应用指南

应用场景分析

实施步骤详解

步骤一：环境准备

conda create -n ai_env python=3.9
conda activate ai_env
pip install numpy pandas matplotlib seaborn scikit-learn tensorflow torch jupyter notebook

步骤二：项目结构

project/
├── data/          # 数据目录
│   ├── raw/       # 原始数据
│   └── processed/ # 处理后数据
├── notebooks/     # Jupyter 笔记本
├── src/           # 源代码
│   ├── data/      # 数据处理
│   ├── features/  # 特征工程
│   ├── models/    # 模型定义
│   └── utils/     # 工具函数
├── tests/         # 测试代码
├── configs/       # 配置文件
├── requirements.txt
└── README.md

步骤三：模型开发流程

最佳实践分享

• 代码规范：使用类型注解、编写文档字符串、遵循 PEP8 规范、添加单元测试。
• 实验管理：使用版本控制、记录实验参数、保存模型检查点、可视化训练过程。

案例分析

房价预测模型

使用机器学习方法预测房屋价格，包含数据预处理、特征工程、模型训练完整流程。

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor
from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error
import matplotlib.pyplot as plt

class HousePricePredictor:
    def __init__(self):
        self.model = None
        self.preprocessor = None

    def prepare_data(self, data: pd.DataFrame, target_col: str):
        X = data.drop(columns=[target_col])
        y = data[target_col]
        numeric_features = X.select_dtypes(include=[np.number]).columns.tolist()
        categorical_features = X.select_dtypes(exclude=[np.number]).columns.tolist()
        self.preprocessor = ColumnTransformer(
            transformers=[
                ('num', StandardScaler(), numeric_features),
                ('cat', OneHotEncoder(handle_unknown='ignore'), categorical_features)
            ]
        )
        return train_test_split(X, y, test_size=0.2, random_state=42)

    def train(self, X_train, y_train):
        self.model = Pipeline([
            ('preprocessor', self.preprocessor),
            ('regressor', GradientBoostingRegressor(n_estimators=200, learning_rate=0.1, max_depth=5, random_state=42))
        ])
        self.model.fit(X_train, y_train)
        return self

    def evaluate(self, X_test, y_test):
        y_pred = self.model.predict(X_test)
        metrics = {
            'RMSE': np.sqrt(mean_squared_error(y_test, y_pred)),
            'MAE': mean_absolute_error(y_test, y_pred),
            'R2': r2_score(y_test, y_pred)
        }
        return metrics, y_pred

    def plot_predictions(self, y_test, y_pred):
        plt.figure(figsize=(10, 6))
        plt.scatter(y_test, y_pred, alpha=0.5)
        plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], 'r--')
        plt.xlabel('真实价格')
        plt.ylabel('预测价格')
        plt.title('房价预测结果')
        plt.show()

if __name__ == "__main__":
    pass

实施效果

过拟合问题解决方案

某模型在训练集表现优秀，但测试集效果很差（泛化能力差）。 改进措施：增加数据量、使用正则化、添加 Dropout、早停法。

常见问题解答

技术问题

Q1：如何选择合适的模型？

Q2：如何处理数据不平衡？

from imblearn.over_sampling import SMOTE
from imblearn.under_sampling import RandomUnderSampler
from sklearn.utils.class_weight import compute_class_weight

# 方法 1：过采样
smote = SMOTE(random_state=42)
X_resampled, y_resampled = smote.fit_resample(X, y)

# 方法 2：欠采样
undersampler = RandomUnderSampler(random_state=42)
X_resampled, y_resampled = undersampler.fit_resample(X, y)

# 方法 3：类别权重
class_weights = compute_class_weight('balanced', classes=np.unique(y), y=y)

应用问题

Q3：如何提升模型性能？ ① 数据增强 ② 特征工程 ③ 模型集成 ④ 超参数调优

Q4：如何避免常见错误？ ⚠️ 注意事项：数据泄露问题、评估方法正确、超参数合理、代码可复现。