Free Engineering Workbench calculator
Regression model error
Measure prediction error for a numeric model: MAE, RMSE, R² and bias.
Data & analyticsHow it's worked out
- RMSE = √(Σe² ÷ n). MAE = Σ|e| ÷ n. R² = 1 − SSE ÷ SST
- Errors from a hold-out set, not training data
NoteRMSE punishes large errors more than MAE. Compare against a naive baseline, such as the mean.
Source: Hyndman and Athanasopoulos, Forecasting: Principles and Practice; standard definitions
Need help acting on this number?
VVnT SeQuor's GenAI & Agentic AI Development practice helps teams turn this metric into a plan.