ex6:SVM

import numpy as np
import matplotlib
import matplotlib.pyplot as plt
from scipy.io import loadmat
from sklearn import svm
from sklearn.datasets import make_classification


data = loadmat('ex6data1.mat')
X = data['X']
Y = data['y']

# linear kernel
def plot_data_desicion(X, y):
    pos = np.where(Y==1)[0]  #表示行索引
    neg = np.where(Y==0)[0]
    plt.figure()
    plt.scatter(X[pos,0], X[pos,1], marker='+', c='red')
    plt.scatter(X[neg,0], X[neg,1], marker='o', c='blue')
    # training
    clf = svm.SVC(kernel='linear', C=100)
    clf.fit(X, Y)
    # accuracy
    print('Test score: %.4f' % clf.score(X, Y))
    # 方向向量
    w = clf.coef_
    b = clf.intercept_
    x = np.linspace(np.min(X[:,0]),np.max(X[:,0]),100)
    y = -(w[0,0]*x+b )/ w[0,1]
    plt.plot(x, y)
    plt.title('Decision Boundary')
    plt.show()


data = loadmat('ex6data2.mat')
X = data['X']
Y = data['y']

# non-linear kernel
def plot_data_desicion(X, Y):
    pos = np.where(Y==1)[0]  #表示行索引
    neg = np.where(Y==0)[0]
    plt.figure()
    plt.scatter(X[pos,0], X[pos,1], marker='+', c='red')
    plt.scatter(X[neg,0], X[neg,1], marker='o', c='blue')
    # training
    clf = svm.SVC(kernel='rbf',gamma=100, C=100)
    clf.fit(X, Y)
    # accuracy
    print('Test score: %.4f' % clf.score(X, Y))
    a = np.transpose(np.linspace(np.min(X[:, 0]), np.max(X[:, 0]), 100).reshape(1, -1))
    b = np.transpose(np.linspace(np.min(X[:, 1]), np.max(X[:, 1]), 100).reshape(1, -1))
    X1, X2 = np.meshgrid(a, b)
    boundary = np.zeros(X1.shape)
    for i in range(X1.shape[1]):
        this_X = np.hstack((X1[:, i].reshape(-1, 1), X2[:, i].reshape(-1, 1)))
        boundary[:, i] = clf.predict(this_X)
    plt.contour(X1, X2, boundary, [0, 1], color='blue')
    plt.title('Decision Boundary')
    plt.show()

plot_data_desicion(X,Y)

#

def plot_data_desicion(X, Y):
    data = loadmat('ex6data3.mat')
    X = data['X']
    Y = data['y']
    Xval = data['Xval']
    Yval = data['yval']

    pos = np.where(Y==1)[0]  #表示行索引
    neg = np.where(Y==0)[0]
    plt.figure()
    plt.scatter(X[pos,0], X[pos,1], marker='+', c='red')
    plt.scatter(X[neg,0], X[neg,1], marker='o', c='blue')
    # training
    steps = [0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30]
    score, this_gamma, this_C = 0, 0, 0
    for C in steps:
        for gamma in steps:
            clf = svm.SVC(kernel='rbf', C=C, gamma=gamma)
            clf.fit(X, np.ravel(Y))
            if clf.score(Xval, np.ravel(Yval)) > score:
                score = clf.score(Xval, np.ravel(Yval))
                this_C = C
                this_gamma = gamma
    # accuracy
    # print('Test score: %.4f' % clf.score(X, Y))
    model =  svm.SVC(kernel='rbf', C=this_C, gamma=this_gamma)
    model.fit(X, np.ravel(Y))
    a = np.transpose(np.linspace(np.min(X[:, 0]), np.max(X[:, 0]), 100).reshape(1, -1))
    b = np.transpose(np.linspace(np.min(X[:, 1]), np.max(X[:, 1]), 100).reshape(1, -1))
    X1, X2 = np.meshgrid(a, b)
    boundary = np.zeros(X1.shape)
    for i in range(X1.shape[1]):
        this_X = np.hstack((X1[:, i].reshape(-1, 1), X2[:, i].reshape(-1, 1)))
        boundary[:, i] = model.predict(this_X)
    plt.contour(X1, X2, boundary, [0, 1], color='blue')
    plt.title('Decision Boundary')
    plt.show()
plot_data_desicion(X,Y)