Mean Square Error (MSE) and Mean Absolute Error (MAE) – Regression
Vishal Chaudhary, PhD
![Vishal Chaudhary, PhD]()
Vishal Chaudhary, PhD
AI/ML Product Leader | Computer Vision & Embedded Vision Specialist | Driving Innovation in Defense, Transport & Intelligent Systems
1. Mean Absolute Error (MAE) – It calculates mean of absolute difference between y and ybar. The equation is defined below
Let’s calculate MAE for an example
Observations:
a. Contribution of difference towards MAE is almost similar
b. Error contribution is almost uniform and hence similar penalization during penalization
c. Outliers have less affect
To be continued with Gradient Descent.....