wmkouw /
libTLDA
| @@ 90-134 (lines=45) @@ | ||
| 87 | # Return transformation matrix and principal component coefficients |
|
| 88 | return V, CX, CZ |
|
| 89 | ||
| 90 | def fit(self, X, y, Z): |
|
| 91 | """ |
|
| 92 | Fit/train a classifier on data mapped onto transfer components. |
|
| 93 | ||
| 94 | INPUT (1) array 'X': source data (N samples by D features) |
|
| 95 | (2) array 'y': source labels (N samples by 1) |
|
| 96 | (3) array 'Z': target data (M samples by D features) |
|
| 97 | OUTPUT None |
|
| 98 | """ |
|
| 99 | # Data shapes |
|
| 100 | N, DX = X.shape |
|
| 101 | M, DZ = Z.shape |
|
| 102 | ||
| 103 | # Assert equivalent dimensionalities |
|
| 104 | assert DX == DZ |
|
| 105 | ||
| 106 | # Transfer component analysis (store target subspace) |
|
| 107 | V, CX, self.CZ = self.subspace_alignment(X, Z, num_components=self. |
|
| 108 | num_components) |
|
| 109 | ||
| 110 | # Map source data onto source principal components |
|
| 111 | X = np.dot(X, CX) |
|
| 112 | ||
| 113 | # Align source data to target subspace |
|
| 114 | X = np.dot(X, V) |
|
| 115 | ||
| 116 | # Train a weighted classifier |
|
| 117 | if self.loss == 'logistic': |
|
| 118 | # Logistic regression model with sample weights |
|
| 119 | self.clf.fit(X, y) |
|
| 120 | elif self.loss == 'quadratic': |
|
| 121 | # Least-squares model with sample weights |
|
| 122 | self.clf.fit(X, y) |
|
| 123 | elif self.loss == 'hinge': |
|
| 124 | # Linear support vector machine with sample weights |
|
| 125 | self.clf.fit(X, y) |
|
| 126 | else: |
|
| 127 | # Other loss functions are not implemented |
|
| 128 | raise NotImplementedError |
|
| 129 | ||
| 130 | # Mark classifier as trained |
|
| 131 | self.is_trained = True |
|
| 132 | ||
| 133 | # Store training data dimensionality |
|
| 134 | self.train_data_dim = DX |
|
| 135 | ||
| 136 | def predict(self, Z_, whiten=False): |
|
| 137 | """ |
|
| @@ 172-209 (lines=38) @@ | ||
| 169 | return np.sum(2*np.clip(1-Xyt[ix], 0, None).T * -Xy[ix, :].T, |
|
| 170 | axis=1).T + np.sum(-4*Xy[~ix, :], axis=0) + 2*l2*theta |
|
| 171 | ||
| 172 | def fit(self, X, y, Z): |
|
| 173 | """ |
|
| 174 | Fit/train an structural correpondence classifier. |
|
| 175 | ||
| 176 | INPUT (1) array 'X': source data (N samples by D features) |
|
| 177 | (2) array 'y': source labels (N samples by 1) |
|
| 178 | (3) array 'Z': target data (M samples by D features) |
|
| 179 | OUTPUT None |
|
| 180 | """ |
|
| 181 | # Data shapes |
|
| 182 | N, DX = X.shape |
|
| 183 | M, DZ = Z.shape |
|
| 184 | ||
| 185 | # Assert equivalent dimensionalities |
|
| 186 | assert DX == DZ |
|
| 187 | ||
| 188 | # Augment features |
|
| 189 | X, _, self.C = self.augment_features(X, Z) |
|
| 190 | ||
| 191 | # Train a classifier |
|
| 192 | if self.loss == 'logistic': |
|
| 193 | # Logistic regression model |
|
| 194 | self.clf.fit(X, y) |
|
| 195 | elif self.loss == 'quadratic': |
|
| 196 | # Least-squares model |
|
| 197 | self.clf.fit(X, y) |
|
| 198 | elif self.loss == 'hinge': |
|
| 199 | # Linear support vector machine |
|
| 200 | self.clf.fit(X, y) |
|
| 201 | else: |
|
| 202 | # Other loss functions are not implemented |
|
| 203 | raise NotImplementedError |
|
| 204 | ||
| 205 | # Mark classifier as trained |
|
| 206 | self.is_trained = True |
|
| 207 | ||
| 208 | # Store training data dimensionality |
|
| 209 | self.train_data_dim = DX + self.num_components |
|
| 210 | ||
| 211 | def predict(self, Z_): |
|
| 212 | """ |
|
