Skip to main content
All CollectionsVRIFY AIVRIFY AI Prediction Maps
Understanding: Input Features
Understanding: Input Features

Learn the key factors involved in setting Features in a VRIFY AI experiment.

Updated over a week ago

Overview

This section is where the features, or data grids, that will be used in the model are defined. The quality and relevance of the chosen features play a crucial role in shaping the model’s performance. Including the appropriate features can improve the model results and overall performance.

Read on for more context and explanations about what this step entails.

Key Concepts by Parameter

Parameter: Features selection

  • The features in the list are compiled from the data provided by the VRIFY AI team.

    • Some geological data layers that have missing data within the AOI rectangle will have some interpolated data. VRIFY will also predict new data based on other relevant data to improve accuracy of rasterization.

    • Point data, like geochemistry sampling points, will be rasterized using an AI-driven interpolation method.

  • Feature file names containing “lowQC” indicate that the interpolation accuracy is lower or that the input data is of lower reliability. These “lowQC” features can still be valuable for analyzing key exploration targets and are included for completeness.

Parameter: Correlation Threshold

  • Setting a correlation threshold helps prevent an imbalance of too much importance on duplication in the dataset by indicating how similar two features must be before one is omitted from the features list.

  • The default threshold is 0.90 (features must be 90% similar to be removed as a duplicate).

    • If the threshold is set to 0, every feature selected will be used, regardless of similarity. If set to 100, only exact duplicate features will be removed.

Parameter: Visualize Input Features and View Input Feature Layer

  • In the Visualize Terrain popout, you can select different features to visualize from your features list. Each layer can be given a colour scheme.

  • Selecting a Terrain Layer allows you to see your topography relative to other geological data layers for comparison.

    • This step is to help validate that the features you are selecting are the ones intended. Setting a terrain layer in this step does not impact the outcomes of your experiment.

  • Use the Vertical Exaggeration slider at the bottom of the screen to adjust the scale of the Z axis for the selected raster layer. This can help you identify subtleties in the rasters more readily.

Notes / Best Practices

  • If you aren’t sure which to include, start by including all available features to see how they behave, then omit any that appear to be skewing the results in later steps.

  • It is good to include the layers that are most relevant to your exploration strategy, like faults, soil anomalies and relevant geophysical surveys. However, the power of VRIFY AI is the ability to test new theories and strategies quickly and easily.

    • Try creating a Prediction Map using, for example, only geophysical surveys or trace element soils. Testing and comparing the results from different data combinations will help you learn more about the project area and mineralization controls.

Still have questions?

Reach out to your dedicated VRIFY AI Contact or email Support@VRIFY.com for more information.

Did this answer your question?