Type 2 Errors
Encord Computer Vision Glossary
In machine learning, a Type 2 error, also known as a false negative, occurs when a model incorrectly predicts the absence of a certain condition or attribute when it is actually present. For example, a medical diagnostic model may fail to detect the presence of a disease in a patient.
Type 2 errors can be a significant problem in machine learning applications where the consequences of false negatives can be costly or harmful. For example, a model that fails to detect fraudulent activity in financial transactions can result in significant financial losses.
To reduce the risk of Type 2 errors in machine learning, various techniques can be employed such as:
- Increasing the model's sensitivity: It is done by decreasing the decision threshold for positive predictions. This can lead to a higher rate of true positives, but may also increase the number of false positives.
- Augment the training data: The training data is augmented with more examples of the less prevalent class. This can help the model to learn the characteristics of the less prevalent class more effectively, leading to a lower rate of false negatives.