geometry_ml_clustering

Geometry ML Clustering

This example demonstrates combining the geometry module with machine learning for 3D spatial clustering and visualization.

What You'll Learn

• Using geometry module for 3D points
• Applying ML clustering to geometric data
• Visualizing 3D clusters
• Combining multiple VSL modules

Prerequisites

• V compiler installed (download here)
• VSL library installed (installation guide)
• No additional system dependencies required

Running the Example

# Navigate to this directory
cd examples/geometry_ml_clustering

# Run the example
v run main.v

Expected Output

The example generates:

• Console output: Data generation, training progress, and accuracy
• 3D Plot: Visualization of clusters with centroids marked

Code Walkthrough

1. Generating Geometric Data

import vsl.gm
import rand

rand.seed([u32(12345), u32(67890)])
mut points := []gm.Point{}

// Create 3D points using geometry module
x, y, z := 1.0, 2.0, 3.0
points << gm.Point.new(x, y, z)

We use the geometry module to create 3D points representing two clusters.

2. Converting to ML Format

import vsl.gm

mut feature_matrix := [][]f64{}
points := []gm.Point{} // Assume populated

// Convert to feature matrix (2D projection for Kmeans)
for p in points {
	feature_matrix << [p.x, p.y]
}

Geometry points are converted to feature vectors for ML algorithms.

3. Clustering and Visualization

We apply K-means clustering and visualize results in 3D space.

Use Cases

• Spatial Analysis: Cluster geographic or spatial data
• Point Cloud Processing: Analyze 3D point clouds
• Robotics: Cluster sensor readings in 3D space
• Computer Vision: Segment 3D objects

Experiment Ideas

Try modifying the example to:

• More clusters: Add a third or fourth cluster
• Different geometries: Use different point distributions
• Distance metrics: Use geometry module's distance functions
• Real data: Import actual 3D point cloud data
• Visualization: Add cluster boundaries or surfaces

Related Examples

• gm_basic_geometry - Basic geometry operations
• ml_kmeans - K-means clustering
• plot_scatter3d_1 - 3D plotting

Related Tutorials

• Geometry Module
• Clustering
• 3D Visualization

Troubleshooting

Low accuracy: Try different initial centroids

Plot doesn't open: Ensure web browser is installed

Module errors: Verify VSL installation

---

Explore more geometry and ML examples in the examples directory.