Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add Gmm sampler #98

Merged
merged 9 commits into from
May 20, 2023
Merged

Add Gmm sampler #98

Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
ee9ca2b
init
mateusz-wozny Mar 8, 2023
bb20635
update docstrings
mateusz-wozny Mar 8, 2023
6e70b95
change typing
mateusz-wozny Mar 15, 2023
263a722
add more tests
mateusz-wozny Apr 23, 2023
693b8ff
add future annotations
mateusz-wozny Apr 23, 2023
91f6b58
add example notebook
mateusz-wozny Apr 24, 2023
e5b8710
black + new code
mateusz-wozny Apr 25, 2023
5da701e
fix docs
mateusz-wozny Apr 27, 2023
aff0caf
merge oversamples at the end
mateusz-wozny May 3, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
merge oversamples at the end
  • Loading branch information
@mateusz-wozny
mateusz-wozny committed May 3, 2023
commit aff0cafe8e6b508e544966dd6a30ffff93eac563
25 changes: 15 additions & 10 deletions multi_imbalance/resampling/gmm_sampler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -354,29 +354,34 @@ def _resample(self, X: np.ndarray, y: np.ndarray) -> Tuple[np.ndarray, np.ndarra
return X, y

def _oversample_each_minority_class(self, X: np.ndarray, y: np.ndarray) -> Tuple[np.ndarray, np.ndarray]:
X_copy = X.copy()
y_copy = y.copy()
X_oversample = []
y_oversample = []
X_oversample.append(X)
y_oversample.append(y)
for minority_class in self.minority_classes:
self.__x_subset = X_copy[y_copy == minority_class]
X, y = self._oversample(X_copy, y_copy, minority_class)
X_copy, y_copy = X, y
self.__x_subset = X[y == minority_class]
X_subset_oversample, y_subset_oversample = self._oversample(X, y, minority_class)
X_oversample.append(X_subset_oversample)
y_oversample.append(y_subset_oversample)
self.__x_subset = None
return X, y

return np.vstack(X_oversample), np.hstack(y_oversample)

def _oversample(self, X: np.ndarray, y: np.ndarray, minority_class: int) -> Tuple[np.ndarray, np.ndarray]:
means, covariances = self._get_coefficients(self.gaussian_mixtures[minority_class])

probabilities = self._get_probas_for_samples_in_component(X, y, minority_class)
quantity_to_generate = self.size_to_align - self.__x_subset.shape[0]

X_subset_oversample = []
y_subset_oversample = []
for component in range(self.gaussian_mixtures[minority_class].n_components):
Nk: np.ndarray = probabilities[component] * quantity_to_generate
x = self._create_samples(means[component], covariances[component], int(Nk))

X = np.append(X, x, axis=0)
y = np.append(y, np.full((x.shape[0],), fill_value=minority_class), axis=0)
X_subset_oversample.append(x)
y_subset_oversample.append([minority_class] * x.shape[0])

return X, y
return np.vstack(X_subset_oversample), np.hstack(y_subset_oversample)

def _get_probas_for_samples_in_component(self, X: np.ndarray, y: np.ndarray, minority_class: int) -> np.ndarray:
X_prob: np.ndarray = self.gaussian_mixtures[minority_class].predict_proba(X[y == minority_class])
Expand Down
Loading