Balanced Accuracy
- Created by
- Renato Passos, Eng. de Software
- Reviewed by
- Renato Passos, Eng. de Software
Last updated: Apr 18, 2026
Formula
BA = (Sens + Esp)/2
About this calculator
The Balanced Accuracy is a measure used in statistics and machine learning to evaluate the performance of classification models. It calculates the average between the sensitivity and specificity of a model, providing a more balanced view of its ability to correctly detect positive and negative classes.
The formula to calculate the Balanced Accuracy is BA = (Sens + Esp)/2, where Sens is the sensitivity (or true positive rate) and Esp is the specificity (or true negative rate). This means that the measure considers both the model's ability to correctly identify positive cases and negative cases.
The Balanced Accuracy is particularly useful in situations where the classes in the dataset are imbalanced, i.e., when one class has many more instances than the other. In these cases, metrics like overall accuracy can be misleading, as a model that always predicts the majority class can have high accuracy but is not useful in practice.
When using the Balanced Accuracy, it is essential to be careful with the interpretation of the results. A high value does not necessarily mean that the model is good in all classes, but rather that it has a good overall performance in both classes. Additionally, it is crucial to consider other metrics and perform complementary analyses for a comprehensive evaluation of the model.
Frequently asked questions
What is Balanced Accuracy?
It is a measure that calculates the average between the sensitivity and specificity of a classification model, providing a balanced view of its performance.
When to use Balanced Accuracy?
Use it when the classes in the dataset are imbalanced and you need a fair measure of the model's performance.
What is the difference between Balanced Accuracy and overall accuracy?
Overall accuracy can be misleading in imbalanced datasets, while Balanced Accuracy provides a more balanced view, considering sensitivity and specificity.
How to interpret a high Balanced Accuracy value?
A high value indicates good overall performance of the model in both classes, but it is essential to consider other metrics for a comprehensive evaluation.
Is Balanced Accuracy useful for binary classification models?
Yes, it is particularly useful for binary classification models, especially in situations of imbalanced classes.