基于sklearn快速实现11种机器学习分类模型的训练与评估
关于作者
作者：小白熊
作者简介：精通python、matlab、c#语言，擅长机器学习，深度学习，机器视觉，目标检测，图像分类，姿态识别，语义分割，路径规划，智能优化算法，数据分析，各类创新融合等等。
联系邮箱：xbx3144@163.com
科研辅导、知识付费答疑、个性化定制以及其他合作需求请联系作者~
一、概述
在数据科学领域，选择合适的机器学习模型至关重要。随着数据的爆炸性增长，各行各业对机器学习的需求也在不断上升。然而，面对众多的模型选择，如何快速而有效地搭建和评估机器学习模型成为了一个挑战。为此，scikit-learn 库提供了丰富的工具和接口，使得构建机器学习模型变得更加便捷。
本文将详细介绍如何使用 scikit-learn 库实现 11 种流行的分类模型。我们将从数据准备开始，逐步进行模型的搭建、训练和评估过程。通过简单的示例，本文将展示如何快速实施这些模型，帮助你更好地理解每种模型的特性和应用场景。小白也能快速上手机器学习！
二、数据准备
首先需要准备一个数据集，这里使用 Iris 数据集作为示例。并使用train_test_split函数按照8：2的比例划分训练集和验证集。
import pandas as pd
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
# 加载数据集
iris = load_iris()
X = iris.data
y = iris.target
# 划分训练集和验证集
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)
三、模型评估指标
准确率 (Accuracy)：表示模型正确预测的样本占总样本的比例，反映模型的整体性能。
精确率 (Precision)：表示被正确预测为正类的样本占所有被预测为正类样本的比例，评估模型在正类预测上的准确性。
召回率 (Recall)：表示被正确预测为正类的样本占所有实际正类样本的比例，评估模型对正类样本的捕捉能力。
F1 值 (F1-Score)：精确率和召回率的调和平均数，综合反映模型在正类预测上的表现，尤其适用于类别不平衡的数据。
四、11种机器学习分类模型
1. 自适应增强 (AdaBoost)
from sklearn.ensemble import AdaBoostClassifier
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
# 初始化模型
ada_model = AdaBoostClassifier()
# 训练模型
ada_model.fit(X_train, y_train)
# 预测
y_train_pred = ada_model.predict(X_train)
y_val_pred = ada_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'AdaBoost:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
2. 人工神经网络 (ANN)
from sklearn.neural_network import MLPClassifier
# 初始化模型
ann_model = MLPClassifier(max_iter=300)
# 训练模型
ann_model.fit(X_train, y_train)
# 预测
y_train_pred = ann_model.predict(X_train)
y_val_pred = ann_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'ANN:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
3. 决策树 (DT)
from sklearn.tree import DecisionTreeClassifier
# 初始化模型
dt_model = DecisionTreeClassifier()
# 训练模型
dt_model.fit(X_train, y_train)
# 预测
y_train_pred = dt_model.predict(X_train)
y_val_pred = dt_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'Decision Tree:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
4. 额外树 (Extra Trees)
from sklearn.ensemble import ExtraTreesClassifier
# 初始化模型
et_model = ExtraTreesClassifier()
# 训练模型
et_model.fit(X_train, y_train)
# 预测
y_train_pred = et_model.predict(X_train)
y_val_pred = et_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'Extra Trees:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
5. 梯度增强机 (GBM)
from sklearn.ensemble import GradientBoostingClassifier
# 初始化模型
gbm_model = GradientBoostingClassifier()
# 训练模型
gbm_model.fit(X_train, y_train)
# 预测
y_train_pred = gbm_model.predict(X_train)
y_val_pred = gbm_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'Gradient Boosting:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
6、k 近邻 (KNN)
from sklearn.neighbors import KNeighborsClassifier
# 初始化模型
knn_model = KNeighborsClassifier()
# 训练模型
knn_model.fit(X_train, y_train)
# 预测
y_train_pred = knn_model.predict(X_train)
y_val_pred = knn_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'KNN:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
7. 轻梯度增强机 (LightGBM)
import lightgbm as lgb
# 初始化模型
lgb_model = lgb.LGBMClassifier()
# 训练模型
lgb_model.fit(X_train, y_train)
# 预测
y_train_pred = lgb_model.predict(X_train)
y_val_pred = lgb_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'LightGBM:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
8. 逻辑回归 (LR)
from sklearn.linear_model import LogisticRegression
# 初始化模型
lr_model = LogisticRegression(max_iter=200)
# 训练模型
lr_model.fit(X_train, y_train)
# 预测
y_train_pred = lr_model.predict(X_train)
y_val_pred = lr_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'Logistic Regression:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
9. 随机森林 (RF)
from sklearn.ensemble import RandomForestClassifier
# 初始化模型
rf_model = RandomForestClassifier(n_estimators=300, random_state=0)
# 训练模型
rf_model.fit(X_train, y_train)
# 预测
y_train_pred = rf_model.predict(X_train)
y_val_pred = rf_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'Random Forest:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
10. 支持向量机 (SVM)
from sklearn.svm import SVC
# 初始化模型
svm_model = SVC(probability=True)
# 训练模型
svm_model.fit(X_train, y_train)
# 预测
y_train_pred = svm_model.predict(X_train)
y_val_pred = svm_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'Support Vector Machine:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
11. 极限梯度增强 (XGBoost)
import xgboost as xgb
# 初始化模型
xgb_model = xgb.XGBClassifier(use_label_encoder=False, verbosity=0)
# 训练模型
xgb_model.fit(X_train, y_train)
# 预测
y_train_pred = xgb_model.predict(X_train)
y_val_pred = xgb_model.predict(X_val)
# 训练集指标
accuracy_train = accuracy_score(y_train, y_train_pred)
precision_train = precision_score(y_train, y_train_pred, average='weighted')
recall_train = recall_score(y_train, y_train_pred, average='weighted')
f1_train = f1_score(y_train, y_train_pred, average='weighted')
# 验证集指标
accuracy_val = accuracy_score(y_val, y_val_pred)
precision_val = precision_score(y_val, y_val_pred, average='weighted')
recall_val = recall_score(y_val, y_val_pred, average='weighted')
f1_val = f1_score(y_val, y_val_pred, average='weighted')
# 打印指标
print(f'XGBoost:')
print(f"训练集 Accuracy: {accuracy_train:.2f}, Precision: {precision_train:.2f}, Recall: {recall_train:.2f}, F1-Score: {f1_train:.2f}")
print(f"验证集 Accuracy: {accuracy_val:.2f}, Precision: {precision_val:.2f}, Recall: {recall_val:.2f}, F1-Score: {f1_val:.2f}")
print("*" * 100)
五、结论
通过以上代码示例，实现了 11 种流行的机器学习模型，并在 Iris 数据集上进行了训练和评估。每个模型的训练集和验证集指标可以在终端轻松查看。这些模型各具特点，适用于不同类型的数据和任务，你可以根据具体需求选择合适的模型进行应用。
除此之外，你可以根据需要进一步扩展这些模型的调优、特征工程或交叉验证等步骤，以提高模型的性能和鲁棒性。
希望这篇文章能帮助你更好地理解和运用 scikit-learn 进行机器学习分类任务！
扫一扫
610
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
