AIExplainer
Machine Learning Intermediate 2 min read

What is minority class?

The class or category with the smallest number of instances in a dataset

In machine learning, a minority class refers to the group or category that has the fewest examples or instances in a dataset. This can make it challenging for algorithms to learn and make accurate predictions about this class.

Think of a minority class like a rare species in a forest. Just as it's harder to study and learn about a rare species because there are fewer examples, a minority class in a dataset can be harder for algorithms to learn from because it has fewer instances.

In a dataset of credit card transactions, the minority class might be fraudulent transactions, which are much less common than legitimate transactions. Accurately predicting fraudulent transactions is crucial, but the imbalance in the dataset can make it challenging.

The minority class is often used in imbalanced classification problems, where one class has a significantly larger number of instances than the others. Techniques such as oversampling the minority class, undersampling the majority class, or using class weights are used to handle this imbalance.

One common misconception is that the minority class is always the class of interest, but this is not always the case. In some problems, the majority class may be the one that is most important to predict accurately.

The concept of minority class has been around since the early days of machine learning, but it has become increasingly important in recent years with the rise of imbalanced classification problems in areas such as finance, healthcare, and cybersecurity.

minority group rare class underrepresented class

Three products for different needs — explore what’s relevant to you.