DORA's input feature scoring has been updated to improve transparency and better reflect modelling suitability in Step 4: Select Input Features. The composite score (0–100) is now the average of four equally weighted criteria instead of five: mineralization relationship, model alignment, raster redundancy, and frequency and sampling compatibility.

You can now see how a raster performed against each individual criterion, not just its overall score. This makes it easier to understand why a layer scored well or poorly and whether to include it based on your geological judgement.

AOI Coverage is no longer included in the composite score calculation. It now appears as a standalone coverage percentage, showing what proportion of your AOI the raster covers. This means a raster is no longer penalized for coverage gaps when its geological relevance is strong.

You can now weigh the two signals independently and decide what matters most for your project.

See How DORA Scores and Ranks Input Features.

You can now export your Learning Points from DORA as a GeoParquet file. This makes it easier to audit your model inputs, share training data with collaborators, or bring your learning points into external GIS tools for further analysis.

See Upload and Manage New Learning Points.

Reach out to your dedicated DORA contact or email support@vrify.com for more information.