For regression models, the evaluation results are based on the following metrics or measures:
- Mean Absolute Error: it tells you how close the predictions are to the actual outcomes.
- Mean Squared Error: it is similar to the absolute error, but it will give a little more weight to the large errors.
- R Squared: it tells you how much better the model’s prediction is than just predicting the mean target value all of the time.
We recommend that you take a look at the evaluations chapter of BigML's Dashboard documentation to learn these formulas and understand how they are computed.