test.py

import SVM
import numpy as np
import timeit
from sklearn.svm import LinearSVC
from sklearn.datasets import load_digits

x = np.random.rand(50000, 1000)
#y = np.zeros((50000,)).astype(int)
y = np.random.choice([-1, 1], size=(50000,))
# for i in range(50000):
#     if np.sum(x[i]) > 2.5:
#         y[i] = 1
#     else:
#         y[i] = -1

# digits = load_digits()
# x = digits.data
# y = digits.target
# for i in range(1797):
#     y[i] = -1 if y[i] == 0 else 1
# print(y)


#x = np.array([[0, 0], [1, 1], [2, 2]]).astype(float)
#y = np.array([-1, 1, 1]).astype(int)

def test():
    svm = SVM.LinearSVM()
    svm.fit(x, y)
    print(svm.get_coef())
    return svm

def svm():
    clf = LinearSVC()
    clf.fit(x, y)
    print("Koef2: ")
    print(clf.coef_)
    return clf


t = timeit.Timer(test)
print(t.timeit(1))
t= timeit.Timer(svm)
print(t.timeit(1))
t = timeit.Timer(test)
print(t.timeit(1))
clf = svm()
a = test()
total = 0
test = np.random.rand(1, 1000)
print(clf.score(x, y))
print(a.score(x, y))
for i in range(9999):
    # test = np.random.rand(1, 5)
    # s = np.sum(test)
    # res1 = clf.predict(test)
    # res2 = a.predict(test)
    # print("S: " + str(s) + ", test: " + str(res2) + ", cor: " + str(res1))
    # clf.predict(test)
    test = np.vstack((test, np.random.rand(1, 1000)))
yGood = clf.predict(test)
yMy = a.predict(test)
print(np.sum(yGood == yMy))