Geometric SMOTE¶

A practical guide¶

One way to fight the imbalanced learning problem is to generate new samples in the classes which are under-represented. Douzas and Bacao (2019) proposed Geometric SMOTE [DB2019]. The GeometricSMOTE class is an implementation of the proposed over-sampling strategy. It offers such scheme:

>>> from collections import Counter
>>> from sklearn.datasets import make_classification
>>> X, y = make_classification(n_classes=3, weights=[0.10, 0.10, 0.80], random_state=0, n_informative=10)
>>> from gsmote import GeometricSMOTE
>>> geometric_smote = GeometricSMOTE()
>>> X_resampled, y_resampled = geometric_smote.fit_resample(X, y)
>>> from collections import Counter
>>> print(sorted(Counter(y_resampled).items()))
[(0, 80), (1, 80), (2, 80)]

The augmented data set should be used instead of the original data set to train a classifier:

>>> from sklearn.tree import DecisionTreeClassifier
>>> clf = DecisionTreeClassifier()
>>> clf.fit(X_resampled, y_resampled) # doctest : +ELLIPSIS
DecisionTreeClassifier(...)

Sample generation¶

Geometric SMOTE represents a modification of the original SMOTE algorithm [CBHK2002] on the data generation mechanism. Considering a sample $x_i$ , a new sample $x_{new}$ will be generated using its k-nearest neighbours (corresponding to k_neighbors). Contrary to SMOTE, instead of generating an artificial observation within a segment between $x_i$ and one of its k-nearest neighbours, G-SMOTE will randomly generate the new artificial observation within a geometry, whose shape is determined by the deformation_factor and truncation_factor. For more information the reader is referred to [DB2019].

Multi-class management¶

GeometricSMOTE can be used with multiple classes as well as binary classes classification. It uses a one-vs-rest approach by selecting each targeted class and computing the necessary statistics against the rest of the data set which are grouped in a single class.

References

[DB2019]

(1, 2) Douzas, G., & Bacao, F. (2019). “Geometric SMOTE: a geometrically enhanced drop-in replacement for SMOTE”, Information Sciences, 501, 118–135. https://doi.org/10.1016/J.INS.2019.06.007

[CBHK2002]

N. V. Chawla, K. W. Bowyer, L. O. Hall, W. P. Kegelmeyer, “SMOTE: synthetic minority over-sampling technique”, Journal of Artificial Intelligence Research, vol. 16, pp. 321-357, 2002.