机器学习之支持向量机

支持向量机

// 机器学习支持向量机
import numpy as np
from sklearn import svm
import matplotlib.pyplot as plt


# 1.	完成数据集的加载和初始化
data = np.loadtxt('bread.txt', delimiter=',')
X, y = np.hsplit(data, [-1, ])
m, n = X.shape

# 2.	将数据集洗牌，合理分割成训练集和测试集
p = np.random.permutation(m)
X, y = X[p], y[p]

size = 0.7
div = int(size*m)
X_train, X_test = np.split(X, [div, ])
y_train, y_test = np.split(y, [div, ])

# 3.	正确调用SVM库函数并训练模型
clf = svm.SVC(C=1, gamma=100)
clf.fit(X_train, y_train.flatten())

# 4.	分别求出训练集和测试集的准确率
h_train = clf.predict(X_train)
print(clf.score(X_train, y_train))
h_test = clf.predict(X_test)
print(clf.score(X_test, y_test))

# 5.	画出整个样本数据并画出分界线
size = 100
x1 = np.linspace(0, 0.8, size)
x2 = np.linspace(0, .5, size)
z = np.zeros((size, size))
for i in range(size):
    for j in range(size):
        z[i][j] = clf.predict(np.array([[x1[i], x2[j]]]))
z = z.transpose()
plt.contour(x1, x2, z)
plt.scatter(X[:, 0], X[:, 1], c=y.ravel())
plt.show()

本文中数据集下载地址:
链接：https://pan.baidu.com/s/1io3aZH-QRCqnGardXOH5yg 提取码：83mu

学习讨论群请加QQ群:521364338,扫码进群领取人工智能学习资料,转载标明出处,侵权必究!