Imbalanced Datasets and Techniques to Handle Them

Class imbalance is a common problem in classification datasets, where the number of data points or observations is not the same across all the classes in the target column. The smaller differences are not a problem. However, there are cases when the dataset has an extreme class imbalance. For example:

Disease screening: We got the dataset to develop a machine learning model that can screen COVID-19 patients. We have only five COVID-19 positive cases in the dataset against 95 COVID-19 negative cases. Say we have 1,000 observations (100 positive and 900 negative cases).

Suppose we train our model on this COVID-19 dataset, and we are happy to see the classifier’s accuracy above 95% with minimal effort. Can we trust the model trained on the dataset with the class distribution of 5:95? It’s an accuracy paradox, where the numbers reflect the underlying class distribution in the imbalanced dataset. Let’s think about it. The baseline accuracy is 95%.

Fraud detection: Another convenient example where only a small fraction of fraudulent cases are present against the fair ones. Sometimes even 1:1000 or 1:5000.

Class imbalance in the dataset can cause frustration and needs to be treated. The following options present the solutions to handle this issue.