What is it about?
Class imbalance is a common problem and it occurs when instances of negative/major class is significantly more than the positive/minor class. The class with scanty instances are always a problem of interest. The conventional learning algorithms are more lenient towards predicting major classes because general rules are more preferred than the specific rules required for predicting the samples from minority class.
Featured Image
Why is it important?
It is observed from the statistical measure that the hybrid over sampled method demonstrated a significant improvement over other combination of the sampling methods for the imbalanced problems.
Perspectives
A hybrid framework is proposed here by pairing three different sampling techniques with a meta-learning algorithm, ‘DECORATE’, which has not been employed so far to study imbalanced problems.
Dr sujata dash
North Orissa University
Read the Original
This page is a summary of: Sampling based hybrid algorithms for imbalanced data classification, International Journal of Hybrid Intelligent Systems, April 2016, IOS Press,
DOI: 10.3233/his-160226.
You can read the full text:
Contributors
The following have contributed to this page
