Model Validation
The Model Validation component provides a set of tools to analyse, compare and assess the consistency of grade distributions across one or more datasets. These datasets may include sample data and/or block model estimates.
When added to point (assay) data, the component can be used to compare the source data with up to three estimation result datasets. This supports validation of estimated grades against the original input information, as well as comparison between estimation outputs.
When added directly to a block model, the component can be used to compare up to three block model result datasets. In this case, the comparison is performed between estimation results only and does not include the source point data.
Model Validation enables the comparison of multiple data sources through a range of statistical summaries and visual diagnostics. It supports the evaluation of how closely datasets align in terms of distribution, trends and global statistics.
A variety of plots and tables are available to facilitate this analysis, including:
- Trend Plots – Model Validation – Used to analyse spatial variation along defined directions, within a series of slices. The slices can be oriented in the X (Northing), Y (Easting) or Z (Elevation) directions.
- Validation Histograms – Used to compare the frequency distribution of values across datasets.
- Validation Probability Plots – Used to compare grade distributions. They can be displayed in four different ways: Cumulative Distribution, Log Cumulative Distribution, Probability and Log Probability. The type of plot is determined by the axis scaling selected.
- Validation QQ Plots – Used to evaluate similarities between datasets by comparing their quantiles. A straight line indicates identical distributions.
- Validation Grade Tonnage Curve – Calculates a grade tonnage curve that can only be inserted below a Model Validation component (and is automatically inserted whenever a new Model Validation component is created).
- Statistical Comparison Tables – Summarises key statistics and differences between datasets.
These tools can be used to:
- Compare distributions between different datasets
- Assess consistency between estimates
- Identify biases or discrepancies
- Evaluate spatial trends and continuity
For a typical workflow focused on the validation of estimated grades, refer to Validate Grade Estimates.
Data used in Model Validation can be weighted according to volume and density where applicable, ensuring that comparisons reflect the relative contribution of each element within the dataset.
