Dataset generation for classification problems Parameters ---------- dataset : str type of classification problem (see code) n : int number of training samples nz : float noise level (>0) p : float proportion of one class in the binary setting
(dataset, n, nz=0.5, theta=0, p=0.5, random_state=None, **kwargs)
| 84 | |
| 85 | |
| 86 | def make_data_classif(dataset, n, nz=0.5, theta=0, p=0.5, random_state=None, **kwargs): |
| 87 | """Dataset generation for classification problems |
| 88 | |
| 89 | Parameters |
| 90 | ---------- |
| 91 | dataset : str |
| 92 | type of classification problem (see code) |
| 93 | n : int |
| 94 | number of training samples |
| 95 | nz : float |
| 96 | noise level (>0) |
| 97 | p : float |
| 98 | proportion of one class in the binary setting |
| 99 | random_state : int, RandomState instance or None, optional (default=None) |
| 100 | If int, random_state is the seed used by the random number generator; |
| 101 | If RandomState instance, random_state is the random number generator; |
| 102 | If None, the random number generator is the RandomState instance used |
| 103 | by `np.random`. |
| 104 | |
| 105 | Returns |
| 106 | ------- |
| 107 | X : ndarray, shape (n, d) |
| 108 | `n` observation of size `d` |
| 109 | y : ndarray, shape (n,) |
| 110 | labels of the samples. |
| 111 | """ |
| 112 | generator = check_random_state(random_state) |
| 113 | |
| 114 | if dataset.lower() == "3gauss": |
| 115 | y = np.floor((np.arange(n) * 1.0 / n * 3)) + 1 |
| 116 | x = np.zeros((n, 2)) |
| 117 | # class 1 |
| 118 | x[y == 1, 0] = -1.0 |
| 119 | x[y == 1, 1] = -1.0 |
| 120 | x[y == 2, 0] = -1.0 |
| 121 | x[y == 2, 1] = 1.0 |
| 122 | x[y == 3, 0] = 1.0 |
| 123 | x[y == 3, 1] = 0 |
| 124 | |
| 125 | x[y != 3, :] += 1.5 * nz * generator.randn(sum(y != 3), 2) |
| 126 | x[y == 3, :] += 2 * nz * generator.randn(sum(y == 3), 2) |
| 127 | |
| 128 | elif dataset.lower() == "3gauss2": |
| 129 | y = np.floor((np.arange(n) * 1.0 / n * 3)) + 1 |
| 130 | x = np.zeros((n, 2)) |
| 131 | y[y == 4] = 3 |
| 132 | # class 1 |
| 133 | x[y == 1, 0] = -2.0 |
| 134 | x[y == 1, 1] = -2.0 |
| 135 | x[y == 2, 0] = -2.0 |
| 136 | x[y == 2, 1] = 2.0 |
| 137 | x[y == 3, 0] = 2.0 |
| 138 | x[y == 3, 1] = 0 |
| 139 | |
| 140 | x[y != 3, :] += nz * generator.randn(sum(y != 3), 2) |
| 141 | x[y == 3, :] += 2 * nz * generator.randn(sum(y == 3), 2) |
| 142 | |
| 143 | elif dataset.lower() == "gaussrot": |