0
$\begingroup$

I have to deal with a small dataset. I thought that I maght take advantage of resamplin methods to enlarge the population and improve the performance of my regression algorithm. I heard about SMOTE, but it is used for classification in imbalanced datasets. Is there any method to create synthetic data of a small size dataset? Thanks.

$\endgroup$
0
$\begingroup$

As suggested by others, use oversampling SMOTE, it will balance your data set and create more examples(based on neighbors) equals to your majority class.

$\endgroup$
0
$\begingroup$

please check out library imbalanced-learn (python). Have you some example of code:

#assuming that you have X and y

from imblearn.over_sampling import SMOTE

smote = SMOTE(ratio='minority')
X_sm, y_sm = smote.fit_sample(X, y)

Documentation:

https://imbalanced-learn.readthedocs.io/en/stable/generated/imblearn.under_sampling.TomekLinks.html

https://imbalanced-learn.readthedocs.io/en/stable/generated/imblearn.over_sampling.SMOTE.html

$\endgroup$
-2
$\begingroup$

You can use Bootstrapped regression.

$\endgroup$
-3
$\begingroup$

You can also try the GAN's to generate some pseudo data.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.