机器学习

1.sklearn

致谢
sklearn库介绍 -人类之奴-2022-知乎

1.1 莺尾花数据集

致谢
探索sklearn | 鸢尾花数据集 -薛定谔的小猫咪-2019-博客园
sklearn iris(鸢尾花)数据集应用 -狼之鸿-2018-博客园
笔记篇二：鸢尾花数据集分类 -一罐趣多多-2022-博客园

原始数据画图

from sklearn import datasets  # 引入数据集,sklearn包含众多数据集
from sklearn.model_selection import train_test_split  # 将数据分为测试集和训练集
from sklearn.neighbors import KNeighborsClassifier  # 利用邻近点方式训练数据
import matplotlib.pyplot as plt


def datasets_demo():
    ###引入数据###
    iris = datasets.load_iris()  # 引入iris鸢尾花数据,iris数据包含4个特征变量
    iris_X = iris.data  # 特征变量
    iris_y = iris.target  # 目标值
    # 利用train_test_split进行将训练集和测试集进行分开，test_size占30%
    X_train, X_test, y_train, y_test = train_test_split(iris_X, iris_y, test_size=0.25)
    # print(y_train)  # 我们看到训练数据的特征值分为3类
    ###训练数据###
    # knn = KNeighborsClassifier()  # 引入训练方法
    # knn.fit(X_train, y_train)  # 进行填充测试数据进行训练
    ###预测数据###
    # print(knn.predict(X_test))  # 预测特征值

    # print(y_test)  # 真实特征值
    return iris


def plot_iris_projection(data, feature_names, target, x_index, y_index):
    for t, marker, c in zip([0, 1, 2], '>ox', 'rgb'):
        plt.scatter(data[target == t, x_index], data[target == t, y_index], marker=marker, c=c)
        plt.xlabel(feature_names[x_index])
        plt.ylabel(feature_names[y_index])
    pass


def main1():
    data_ = datasets_demo()
    data = data_['data']  # 为numpy.ndarray类型
    feature_names = data_['feature_names']
    target = data_['target']
    seq = [(0, 1), (0, 2), (0, 3), (1, 2), (1, 3), (2, 3)]
    for i, (x_index, y_index) in enumerate(seq):
        plt.subplot(2, 3, i+1)
        plot_iris_projection(data, feature_names, target, x_index, y_index)
        pass
    plt.show()
    pass
    
main1()

训练测试

import numpy as np
from sklearn import datasets  # 引入数据集,sklearn包含众多数据集
from sklearn.model_selection import train_test_split  # 将数据分为测试集和训练集
from sklearn.linear_model import LogisticRegression
import matplotlib.pyplot as plt

def load_data():
    # 共150条数据，训练120条，测试30条，进行2,8分进行模型训练
    # 每条数据类型为 x{nbarray} [6.4, 3.1, 5.5, 1.8]
    inputdata = datasets.load_iris()
    # 切分，测试训练2,8分
    x_train, x_test, y_train, y_test = \
        train_test_split(inputdata.data, inputdata.target, test_size=0.2, random_state=1)
    return x_train, x_test, y_train, y_test


def main():
    # 训练集x ,测试集x,训练集label,测试集label
    x_train, x_test, y_train, y_test = load_data()
    # l2为正则项
    model = LogisticRegression(penalty='l2', max_iter=10000)  # 引入训练方法
    model.fit(x_train, y_train)  # 进行填充测试数据进行训练

    print("w: ", model.coef_)  # “斜率”参数（w，也叫作权重或系数）被保存在 coef_ 属性中
    print("b: ", model.intercept_)  # 偏移或截距（b）被保存在 intercept_ 属性中
    # 准确率
    print("precision: ", model.score(x_test, y_test))  # 1-残差平方和/总误差平方和 (若残差大，则数据预测不准)
    print("MSE: ", np.mean((model.predict(x_test) - y_test) ** 2))  # 均方误差(Mean squared error)
    pass

main()

画出预测分类结果

因为难以画出4个特征分割的图，因此这里截取2个特征进行训练，画出的分类图也就容易展现了。

def test_plot():
    # 载入数据集
    iris = datasets.load_iris()
    x_train, x_test, y_train, y_test = train_test_split(iris.data[:, :2], iris.target, test_size=0.2, random_state=1)
    X = x_train  # 获取花卉两列数据集
    Y = y_train
    # 逻辑回归模型
    lr = LogisticRegression(C=1e5)  # c=1e5是目标函数
    lr.fit(X, Y)

    # meshgrid函数生成两个网格矩阵
    h = .02
    # 初始化逻辑回归模型并进行训练
    x_min, x_max = X[:, 0].min() - .5, X[:, 0].max() + .5
    y_min, y_max = X[:, 1].min() - .5, X[:, 1].max() + .5
    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))
    # pcolormesh函数将xx,yy两个网格矩阵和对应的预测结果Z绘制在图片上
    Z = lr.predict(np.c_[xx.ravel(), yy.ravel()])
    Z = Z.reshape(xx.shape)
    plt.figure(1, figsize=(8, 6))
    plt.pcolormesh(xx, yy, Z, cmap=plt.cm.Paired)
    plot_iris_projection(X, iris['feature_names'], Y, 0, 1)
    plt.legend(iris['target_names'], loc=2)

    plt.xlim(xx.min(), xx.max())
    plt.ylim(yy.min(), yy.max())
    plt.xticks(())
    plt.yticks(())
    plt.show()

    print("precision: ", lr.score(x_test, y_test))  # 1-残差平方和/总误差平方和 (若残差大，则数据预测不准)
    print("MSE: ", np.mean((lr.predict(x_test) - y_test) ** 2))  # 均方误差(Mean squared error)
    pass
    
test_plot()

结果如图：