README.md

phpOptics

PHP Implementation of the OPTICS Clustering Algorithm

Features

• Direct access to ordered values for custom extraction
• DBSCAN and inflection based cluster extraction included
• Euclidean, Great Circle, and CIEDE2000 Color Distance Included
• Easy to extention for different point distances
• Depending on chosen distance type, easily handles 1000 points (500 is suggested for Color)

References

• Wikipedia OPTICS
• OPTICS: Ordering Points To Identify the Clustering Structure By Mihael Ankerst, Markus M. Breunig, Hans-Peter Kriegel, J&g Sander (pdf)
• CIEDE2000 Color Distance

Basic Usage

$optics = new Optics($epsilonMaxRadius, $minPoints, $pointDimensions = 2, $pointClass = 'OpticsPoint');
$optics->addDataPoints(array(array(1,3,4), ...));
$optics->run();
$optics->getPointsOrdered();
OpticsCluster::extractDBSCANClusters($optics, $clusterEpsilon);