1.KNN
查找距离已知的几个点最近的类型，并返回这个类型进行预测。
如小明在北京，小红在北京，小刚在河南，而我距离小明和小红比小刚近，则我最可能在北京而不是河南
#!/usr/bin/env python # -*- coding: utf-8 -*- # @File : KNN近邻算法.py # @Author: 赵路仓 # @Date : 2020/4/2 # @Desc : 学习网站：https://www.bilibili.com/video/BV1nt411r7tj?p=21 # @Contact : 398333404@qq.com from sklearn.datasets import load_iris from sklearn.model_selection import train_test_split, GridSearchCV from sklearn.preprocessing import StandardScaler from sklearn.neighbors import KNeighborsClassifier import numpy as np def knn_iris(): """ 用KNN算法对鸢尾花进行分类 :return: """ # 1.获取数据 iris = load_iris() print(iris) # 2.划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=6) # 3.特征工程：标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4.KNN算法预估器 estimator = KNeighborsClassifier(n_neighbors=6) estimator.fit(x_train, y_train) # 5.模型评估 # 方法一：直接对比真实数据和预测值 y_predit = estimator.predict(x_test) print("y_predit: ", y_predit) print("对比真实值和预测值: ", y_test == y_predit) # 方法2：计算准确率 score = estimator.score(x_test, y_test) print("准确率为： ", score) # 预测新的鸾尾花品种 x_new = np.array([[5, 2.9, 1, 0.2]]) prediction = estimator.predict(x_new) print(prediction) return None def knn_iris_gscv(): """ 用KNN算法对鸢尾花进行分类,添加网格搜索与交叉验证 :return: """ # 1.获取数据 iris = load_iris() print(iris) # 2.划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=6) # 3.特征工程：标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4.KNN算法预估器 estimator = KNeighborsClassifier(n_neighbors=5) # 加入网格搜索与交叉验证 # 参数准备 从下侧中取n_neighbors param_dict = { "n_neighbors": [1, 3, 5, 7, 9, 11] } estimator = GridSearchCV(estimator, param_grid=param_dict, cv=10) estimator.fit(x_train, y_train) # 5.模型评估 # 方法一：直接对比真实数据和预测值 y_predit = estimator.predict(x_test) print("y_predit: ", y_predit) print("对比真实值和预测值: ", y_test == y_predit) # 方法2：计算准确率 score = estimator.score(x_test, y_test) print("准确率为： ", score) """ 最佳参数：best_params_ 最佳结果：best_score_ 最佳估计器：best_estimator_ 交叉验证结果：cv_results_ """ print("最佳参数: ", estimator.best_params_) print("最佳结果: ", estimator.best_score_) print("最佳估计器: ", estimator.best_estimator_) print("交叉验证结果: ", estimator.cv_results_) # 预测新的鸾尾花品种 x_new = np.array([[5, 2.9, 1, 0.2]]) prediction = estimator.predict(x_new) print(prediction) return None if __name__ == "__main__": # 代码1：KNN对鸾尾花分类 # knn_iris() # 代码2：KNN预测鸾尾花分类并添加网格搜索和交叉验证 knn_iris_gscv()
2.决策树
分类树（决策树）是一种十分常用的分类方法。他是一种监管学习，所谓监管学习就是给定一堆样本，每个样本都有一组属性和一个类别，这些类别是事先确定的，那么通过学习得到一个分类器，这个分类器能够对新出现的对象给出正确的分类。这样的机器学习就被称之为监督学习。
#!/usr/bin/env python # -*- coding: utf-8 -*- # @File : 决策树.py # @Author: 赵路仓 # @Date : 2020/4/3 # @Desc : https://www.bilibili.com/video/BV1nt411r7tj?p=28 # @Contact : 398333404@qq.com import os from sklearn.datasets import load_iris from sklearn.model_selection import train_test_split from sklearn.tree import DecisionTreeClassifier, export_graphviz import graphviz def decision_iris(): """ 用决策树对鸢尾花数据进行分类 :return: """ # 1.获取数据集 iris = load_iris() print(iris.data[1]) print(iris.target[1]) # 2.划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=22) print(y_train) # 3.决策树预估器 estimator = DecisionTreeClassifier(criterion="entropy") estimator.fit(x_train, y_train) # 4.模型评估 # 方法一：直接对比真实数据和预测值 y_predit = estimator.predict(x_test) print("y_predit: ", y_predit) print("对比真实值和预测值: ", y_test == y_predit) # 方法2：计算准确率 score = estimator.score(x_test, y_test) print("准确率为： ", score) # 可视化决策树 # 生成文件 dot_data = export_graphviz(estimator, out_file=None) graph = graphviz.Source(dot_data) graph.render("tree") # tree3是我想要命名的pdf名称 return None if __name__ == "__main__": decision_iris()
3.线性回归
线性回归的任务是找到一个从特征空间X到输出空间Y的最优的线性映射函数
#!/usr/bin/env python # -*- coding: utf-8 -*- # @File : 波士顿房价预测.py # @Author: 赵路仓 # @Date : 2020/4/11 # @Desc : # @Contact : 398333404@qq.com from sklearn.datasets import load_boston from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler from sklearn.linear_model import LinearRegression, SGDRegressor, Ridge from sklearn.metrics import mean_squared_error # 正规方程 def linear1(): """ 正规方程的优化方法对波士顿房价进行预测 :return: """ # 1.获取数据 boston = load_boston() # 2.划分数据集 x_train, x_test, y_train, y_test = train_test_split(boston.data, boston.target, random_state=22) # 3.标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4.预估器 正规方程优化 小于十万条 estimator = LinearRegression() estimator.fit(x_train, y_train) # 5.得出模型 print("正规方程权重系数为：", estimator.coef_) print("正规方程偏置：", estimator.intercept_) # 6.模型评估 y_predit = estimator.predict(x_test) print("预测房价：", y_predit) error = mean_squared_error(y_test, y_predit) print("正规方程-均方误差：", error) return None # 梯度下降 def linear2(): """ 梯度下降的优化方法对波士顿房价进行预测 :return: """ # 1.获取数据 boston = load_boston() # 2.划分数据集 x_train, x_test, y_train, y_test = train_test_split(boston.data, boston.target, random_state=22) # 3.标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4.预估器 梯度下降，eta0学习率，max_iter迭代次数，大量数据推荐使用 estimator = SGDRegressor(learning_rate="constant", eta0=0.001, max_iter=10000) estimator.fit(x_train, y_train) # 5.得出模型 print("梯度下降权重系数为：", estimator.coef_) print("梯度下降偏置：", estimator.intercept_) # 6.模型评估 y_predit = estimator.predict(x_test) print("预测房价：", y_predit) error = mean_squared_error(y_test, y_predit) print("梯度下降-均方误差：", error) return None # 岭回归 def linear3(): """ 岭回归对波士顿房价进行预测 :return: """ # 1.获取数据 boston = load_boston() # 2.划分数据集 x_train, x_test, y_train, y_test = train_test_split(boston.data, boston.target, random_state=22) # 3.标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4.预估器 梯度下降，eta0学习率，max_iter迭代次数，大量数据推荐使用 estimator = Ridge(max_iter=10000) estimator.fit(x_train, y_train) # 5.得出模型 print("岭回归权重系数为：", estimator.coef_) print("岭回归偏置：", estimator.intercept_) # 6.模型评估 y_predit = estimator.predict(x_test) print("预测房价：", y_predit) error = mean_squared_error(y_test, y_predit) print("岭回归-均方误差：", error) return None if __name__ == "__main__": # 代码1：正规方程 linear1() # 代码2：梯度下降 linear2() # 代码3：岭回归 linear3()
4.逻辑回归
简单来说， 逻辑回归（Logistic Regression）是一种用于解决二分类（0 or 1）问题的机器学习方法，用于估计某种事物的可能性。比如某用户购买某商品的可能性，某病人患有某种疾病的可能性，以及某广告被用户点击的可能性等。
#!/usr/bin/env python # -*- coding: utf-8 -*- # @File : 癌症逻辑回归.py # @Author: 赵路仓 # @Date : 2020/4/11 # @Desc : # @Contact : 398333404@qq.com from sklearn.datasets import load_breast_cancer from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler from sklearn.linear_model import LogisticRegression from sklearn.metrics import classification_report, roc_auc_score import pandas as pd import numpy as np def cancer_demo(): """ 利用逻辑回归对乳腺癌进行二分类 :return: """ # 载入数据 cancer = load_breast_cancer() # print(cancer.feature_names) # print(cancer.data) # print(cancer.target) # 划分数据集 x_train, x_test, y_train, y_test = train_test_split(cancer.data, cancer.target) # 标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) print(x_train) # 构建预估器 estimator = LogisticRegression() estimator.fit(x_train, y_train) # 得出模型 print("逻辑回归权重系数：", estimator.coef_) print("逻辑回归偏置：", estimator.intercept_) # 模型评估 # 方法一：直接对比真实数据和预测值 y_predit = estimator.predict(x_test) print("y_predit: ", y_predit) print("对比真实值和预测值: ", y_test == y_predit) # 方法2：计算准确率 score = estimator.score(x_test, y_test) print("准确率为： ", score) # 查看精确率 召回率 以及F1-score report = classification_report(y_test, y_predit, labels=[0, 1], target_names=['良性', '恶性']) print(report) roc=roc_auc_score(y_test,y_predit) print("ROC曲线：",roc) if __name__ == "__main__": cancer_demo()