sklearn各分类模型的比较

在这里插入图片描述
import numpy as np, matplotlib.pyplot as mp
from sklearn.model_selection import train_test_split  # 数据切分
from sklearn.preprocessing import StandardScaler  # 数据标准化
from sklearn.datasets import make_moons, make_circles, make_classification  # 数据集
from sklearn.neural_network import MLPClassifier  # 神经网络
from sklearn.neighbors import KNeighborsClassifier  # K最近邻
from sklearn.svm import SVC  # 支持向量机
from sklearn.gaussian_process import GaussianProcessClassifier  # 高斯过程
from sklearn.gaussian_process.kernels import RBF  # 高斯核函数
from sklearn.tree import DecisionTreeClassifier  # 决策树
from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier  # 集成方法
from sklearn.naive_bayes import GaussianNB  # 朴素贝叶斯
from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis  # 判别分析

# 建模、设定参数
names = ["Nearest Neighbors", "Linear SVM", "RBF SVM", "Gaussian Process", "Decision Tree",
         "Random Forest", "Neural Net", "AdaBoost", "Naive Bayes", "QDA"]
classifiers = [
    KNeighborsClassifier(3),  # K最近邻
    SVC(kernel="linear", C=0.025),  # 线性的支持向量机
    SVC(gamma=2, C=1),  # 径向基函数的支持向量机
    GaussianProcessClassifier(1.0 * RBF(1.0)),  # 基于拉普拉斯近似的高斯过程
    DecisionTreeClassifier(max_depth=5),  # 决策树
    RandomForestClassifier(max_depth=5, n_estimators=10, max_features=1),  # 随机森林
    MLPClassifier(alpha=1),  # 多层感知机
    AdaBoostClassifier(),  # 通过迭代弱分类器而产生最终的强分类器的算法
    GaussianNB(),  # 朴素贝叶斯
    QuadraticDiscriminantAnalysis()]  # 二次判别分析

# 创建随机样本集
X, y = make_classification(n_features=2, n_redundant=0, n_informative=2,
                           random_state=1, n_clusters_per_class=1)
rng = np.random.RandomState(2)
X += 2 * rng.uniform(size=X.shape)
linearly_separable = (X, y)
datasets = [make_moons(noise=0.3, random_state=0),
            make_circles(noise=0.2, factor=0.5, random_state=1),
            linearly_separable]

# 遍历样本集
figure = mp.figure(figsize=(20, 4))
i = 1  # 子图参数
h = .02  # 网眼步长（绘制等高线图的参数）
for ds_cnt, ds in enumerate(datasets):
    # 数据预处理，切分训练集和测试集
    X, y = ds
    X = StandardScaler().fit_transform(X)
    X_train, X_test, y_train, y_test = \
        train_test_split(X, y, test_size=.4, random_state=42)

    # 用于绘制等高线图
    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))

    # 绘制原始样本集
    ax = mp.subplot(len(datasets), len(classifiers) + 1, i)
    if ds_cnt == 0:
        ax.set_title("Input data", size=10)
    ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train,
               edgecolors='k')  # 绘制训练集散点图
    ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test, alpha=0.5,
               edgecolors='k')  # 绘制测试集散点图
    ax.set_xlim(xx.min(), xx.max())
    ax.set_ylim(yy.min(), yy.max())
    ax.set_xticks(())
    ax.set_yticks(())
    i += 1

    # 遍历分类器
    for name, clf in zip(names, classifiers):
        ax = mp.subplot(len(datasets), len(classifiers) + 1, i)
        clf.fit(X_train, y_train)  # 训练
        score = clf.score(X_test, y_test)  # 模型评分

        # 绘制决策边界
        if hasattr(clf, "decision_function"):
            Z = clf.decision_function(np.c_[xx.ravel(), yy.ravel()])
        else:
            Z = clf.predict_proba(np.c_[xx.ravel(), yy.ravel()])[:, 1]
        Z = Z.reshape(xx.shape)
        ax.contourf(xx, yy, Z, 2, alpha=.8)  # 等高线图
        # 散点图
        ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train,
                   edgecolors='k')  # 训练集
        ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test,
                   edgecolors='k', alpha=0.6)  # 测试集
        # 刻度、标签、标题等
        ax.set_xlim(xx.min(), xx.max())
        ax.set_ylim(yy.min(), yy.max())
        ax.set_xticks(())
        ax.set_yticks(())
        if ds_cnt == 0:
            ax.set_title(name, size=10)
        ax.text(xx.max() - .3, yy.min() + .3, ('%.2f' % score).lstrip('0'),
                size=9, horizontalalignment='right')
        i += 1
mp.tight_layout()
mp.show()
En	Cn
boost	n. 推动；vt. 促进
preprocess	vt. 预处理
gamma	希腊语的第三个字母： $\gamma$
quadratic	二次方程式；二次的
discriminant	判别式
percept	n. 认知
radial	n. 射线，光线；adj. 半径的；放射状的
MLP	Multi Layered Perceptron
RBF	Radial Basis Function
linearly separable	线性可分
tick	n. 滴答声；记号；
sklearn各分类模型的比较

猜你喜欢
