# Resource Model Validation AI Agent Connect

> Validate mineral resource block models using statistical analysis and spatial swath plots.

## Overview
- **Category:** data-analysis
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_iZvzoUDDLKeVTunIMJpIhGxIDj9sJ3irnWTaI1wJ/ai-agent-connect
- **Tags:** resource-modeling, block-model, validation, geology, statistics

## Description

This MCP server provides a specialized toolset for evaluating the accuracy and reliability of mineral resource block models. It allows for direct comparison between estimated model grades and measured sample data. Users can perform global statistical comparisons using `get_statistical_summary`, evaluate spatial accuracy through `generate_swath_analysis`, compute advanced error metrics with `calculate_validation_metrics`, and pinpoint geographic discrepancies using `detect_local_bias`.

## Tools

### calculate_validation_metrics
Computes advanced error metrics to quantify model quality

### detect_local_bias
Identifies specific geographic areas where the model is significantly overestimating or underestimating

### generate_swath_analysis
Evaluates the spatial accuracy of the model by comparing averages along a chosen axis

### get_statistical_summary
Provides a high-level comparison of the globalThis averages between the model and the samples

## Prompt Examples

**Prompt:** 
```
Compare the global averages of these sample grades [12.5, 13.2, 11.8] and model grades [12.8, 13.0, 12.1].
```

**Response:** 
```
The model mean is 12.63 and the sample mean is 12.5. The mean difference is -0.13, indicating a slight overestimation by the model.
```

**Prompt:** 
```
Calculate the validation metrics for sample grades [10.0, 15.0] and model grades [11.0, 14.0].
```

**Response:** 
```
The RMSE is 1.0, the mean error is 0.0, and the relative error is 0.0.
```

**Prompt:** 
```
Run a swath analysis along the Z axis with 5 slices using the provided points.
```

**Response:** 
```
The swath analysis along the Z axis is complete. The maximum deviation across the 5 slices is 0.45 grade units.
```

## Frequently Asked Questions

**How can I check for systematic errors in my model?**
You can use `calculate_validation_metrics` to compute RMSE and mean error, or `detect_local_bias` to find specific geographic zones where the model deviates from sample data.

**What is the purpose of swath analysis?**
The `generate_swath_analysis` tool evaluates spatial accuracy by comparing average grades along a chosen axis (X, Y, or Z) within defined slices.

**Does this tool handle sampling bias?**
Yes, the `get_statistical_summary` tool includes an option to apply declustering weights to correct for sampling bias.