Engineers using ModelMill to train Logic-Based Networks for time-series forecasting can now review model accuracy directly from the Training Results page, without leaving the platform or piping data into external tooling.
What's new
The Training Results page now includes two new blocks: the Forecast Block and Per Feature Metrics.
The Forecast Block shows predicted values for your target variable, plotted against your historical data. It gives you a quick read on how the model's output tracks the real series before you commit to deployment.
Per Feature Metrics breaks down accuracy at the feature level. For each LBN trained against a forecasting objective, it shows:
- MAE (Mean Absolute Error) — the average size of prediction errors across the evaluation window
- MAPE (Mean Absolute Percentage Error) — percentage-based error, useful when you're comparing models trained on series with different scales
- RMSE (Root Mean Squared Error) — weights larger errors more heavily, so outlier performance shows up clearly
Who this is for
Any engineer training LBNs against a forecasting or time-series prediction objective. If your training session includes at least one model with a forecasting target, both blocks appear automatically in Training Results.
How to use it
- Open a completed training session from the Training Results page.
- Scroll to the Forecast Block and Per Feature Metrics — both appear below the primary results summary.
- Select the model you want to evaluate from the model selector at the top of the block.
- Check the prediction chart in the Forecast Block, then review the per-feature accuracy figures.
- Use the Compare toggle to put results from multiple trained models side by side.
Notes
Results are saved alongside the training session and can be revisited at any time from Training History. Historical data is shown separately in the Historical Data block.
The old version sends its regards.