SVM是一个很庞杂的体系,前面我从以下几个方面分别讨论了SVM的大致原理:
- 机器学习之SVM(Hinge Loss+Kernel Trick)原理推导与解析
- 强对偶性、弱对偶性以及KKT条件的证明(对偶问题的几何证明)
- 核方法概述----正定核以及核技巧(Gram矩阵推导正定核)
- 手推序列最小优化(sequential minimal optimization,SMO)算法
- 再谈SVM(hard-margin和soft-margin详细推导、KKT条件、核技巧)
本篇博文主要是对SVM系列博客的一个实践,手写SVM来简单地对指定数据集进行分类,数据及完整代码放在了最后!
import pandas as pd
import numpy as np
from sklearn import svm
import matplotlib.pyplot as plt
def SVM():
C = 0.6
toler = 0.001
kernel = ['linear', 0.0]
model, lambdas, w, b = smo_main(C, kernel, toler)
x, y, train_x, train_y, test_x, test_y = load_data('SVM数据集/testSet.txt')
# test_x, test_y = load_data('horse_colic/horseColicTest.txt')
sum = 0
for i in range(len(test_y)):
res = np.dot(w.T, test_x[i, :]) + b
res = np.sign(res)
if res == test_y[i]:
sum += 1
print('手写正确率:%.2f%%' % (sum / len(test_y) * 100))
plotSVM(model, w)
def sklearn_svm():
x, y, train_x, train_y, test_x, test_y = load_data('SVM数据集/testSet.txt')
clf = svm.SVC()
clf.fit(train_x, train_y)
print('调包正确率:%.2f%%' % (clf.score(test_x, test_y) * 100))
if __name__ == '__main__':
SVM()
sklearn_svm()
结果展示:
完整代码放在了GitHub,下载时麻烦给个star!!谢谢!!
源码