Python西瓜书使用数据集3.0α线性核和高斯核训练SVM+散点图可视化

西瓜数据集3.0α

# -*- coding: utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
from sklearn import svm
import pandas as pd
from sklearn.metrics import accuracy_score#返回正确的比例
from sklearn.preprocessing import LabelEncoderplt.rcParams[&＃39;font.sans-serif&＃39;] &＃61; [&＃39;SimHei&＃39;] #用来正常显示中文标签
plt.rcParams[&＃39;axes.unicode_minus&＃39;] &＃61; False #用来正常显示负号
plt.close(&＃39;all&＃39;)def main():
#1.获取x,ydata &＃61; pd.read_table(&＃39;watermelon30a.txt&＃39;,delimiter&＃61;&＃39;,&＃39;) x &＃61; pd.DataFrame({&＃39;密度&＃39;:data[&＃39;密度&＃39;],&＃39;含糖率&＃39;:data[&＃39;含糖率&＃39;]})x &＃61; x.values.tolist() encoder &＃61; LabelEncoder()#将好瓜坏瓜映射为1/0y &＃61; encoder.fit_transform(data[&＃39;好瓜&＃39;]).tolist()x,y &＃61; np.array(x),np.array(y)
#2.1.线性核 linear_svm &＃61; svm.SVC(C&＃61;0.5, #惩罚参数kernel&＃61;&＃39;linear&＃39;)linear_svm.fit(x,y)y_pred &＃61; linear_svm.predict(x)print(&＃39;**linear_svm的准确率**: %s&＃39; %(accuracy_score(y_pred&＃61;y_pred, y_true&＃61;y)))
##2.2.高斯核gauss_svm &＃61; svm.SVC(C&＃61;0.5,kernel&＃61;&＃39;rbf&＃39;)gauss_svm.fit(x,y)y_pred2 &＃61; gauss_svm.predict(x)print(&＃39;**gauss_svm的准确率**: %s&＃39; %(accuracy_score(y_pred&＃61;y_pred2, y_true&＃61;y))) class_method &＃61; {&＃39;线性核&＃39;:linear_svm,&＃39;高斯核&＃39;:gauss_svm}visual(data,class_method)##数据特征可视化
def visual(data,class_method):colormap &＃61; dict(zip(data[&＃39;好瓜&＃39;].value_counts().index.tolist(),[&＃39;blue&＃39;,&＃39;green&＃39;]))#坏瓜好瓜颜色die &＃61; data.groupby(&＃39;好瓜&＃39;) plt.figure()for species,klass in die:plt.scatter(klass[&＃39;密度&＃39;],klass[&＃39;含糖率&＃39;],color &＃61; colormap[species],label &＃61; species)for name,model in class_method.items():sv &＃61; model.support_vectors_plt.plot(sv[:,0],sv[:,1],label&＃61;str(name)&＃43;&＃39;_supported_vector&＃39;) plt.legend(frameon&＃61;True, title&＃61;&＃39;好瓜&＃39;,loc&＃61;"upper left") plt.title(&＃39;SVC&＃39;)plt.show()if __name__&＃61;&＃61;"__main__":main()

结果表明&＃xff0c;使用线性核和高斯训练核的支持向量实际是一样的(两条线重合):