DBSCAN

Introduction

Grid based DBSCAN algorithm of different versions

Algorithms

1. Brute force (use distance matrix)
2. KD Tree based algorithm (KDD96 Paper)
3. Grid based (by Ade Gunawan)
4. Reduced precision grid based (approximate DBSCAN with controllable precision)
5. LSH grid based algorithm (use locality sensitive hashing to merge clusters in any dimension)

Compile & run

$cd CPUCode
$cmake CMakeList.txt
$make
$./test_dbscan

Data & result

1. Sample data is in data/
2. Two results are in result/

Dependencies

Platform and libraries

1. g++/icpc (c++11 or c++0x supported, see CMakeList.txt for details)
2. boost (boost/numeric/ublas)
3. cmake