Positive Class
Positive Class refers to the category representing the condition, event, or outcome that a machine learning model is designed to detect or identify. It is commonly used in binary classification alongside the negative class.
What is Positive Class?
In a binary classification problem, each observation is assigned to either a positive or negative class. For example, in fraud detection, a fraudulent transaction may be defined as the positive class, while a legitimate transaction is the negative class. The meaning of the positive class depends on how the classification task is defined.
Why is Positive Class Important?
Clearly defining the positive class is essential for interpreting model predictions and evaluation metrics such as precision, recall, false positives, and false negatives. It also ensures that training and evaluation datasets use consistent labels.
Common use cases
Positive classes are commonly used in fraud detection, spam detection, medical diagnosis, content moderation, anomaly detection, and other binary classification tasks.