()
| 27 | return X1, y1, X2, y2 |
| 28 | |
| 29 | def gen_lin_separable_overlap_data(): |
| 30 | # generate training data in the 2-d case |
| 31 | mean1 = np.array([0, 2]) |
| 32 | mean2 = np.array([2, 0]) |
| 33 | cov = np.array([[1.5, 1.0], [1.0, 1.5]]) |
| 34 | X1 = np.random.multivariate_normal(mean1, cov, 100) |
| 35 | y1 = np.ones(len(X1)) |
| 36 | X2 = np.random.multivariate_normal(mean2, cov, 100) |
| 37 | y2 = np.ones(len(X2)) * -1 |
| 38 | return X1, y1, X2, y2 |
| 39 | |
| 40 | def split_train(X1, y1, X2, y2): |
| 41 | X1_train = X1[:90] |