MAE MSE. How to understand accuracy?

Question

Neural networks

Hello! After training the linear regression model, I received the following indicators:
MSE: 4903489.792
MAE: 13.217 Tell me

how to interpret them? Is this normal or is it really bad?
The features by which the data is predicted are only dummy variables. Predicted data - percentage (value from 0 to 1)

Answer 1

MAE is easiest to interpret: on average, the model error jumps by +/- 13.2
If you have values ​​​​from 0 to 1, then probably 13.2 is a fairly large spread.
MSE itself is trying to interpret the idea as a whole is such itself, but your example shows that something is wrong, but it is better to take RMSE