# DBSCAN

**Repository Path**: leox24/DBSCAN

## Basic Information

- **Project Name**: DBSCAN
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-09-08
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# DBSCAN
DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is a data clustering algorithm proposed by Martin Ester, Hans-Peter Kriegel, Jörg Sander and Xiaowei Xu in 1996. The algorithm had implemented with pseudocode described in [wiki](https://en.wikipedia.org/wiki/DBSCAN), but it is not optimised.
  
# Example  
You can test this DBSCAN algorithm with example code(main.cpp) & sample data(benchmark_hepta.dat).
  
# Results
Clustering performance was tesed via clustering-benchmark dataset and real-world dataset.  
  
## benchmark dataset  
Artificial dataset was used to test clustering algorithm which can be obtained [here](https://github.com/deric/clustering-benchmark). Following parameters were used:  
1. Minimum number of points: 4
2. Epsilon: 0.75
  
  
![dbscan_benchmark1](https://user-images.githubusercontent.com/23667624/38789657-4e5b7738-412b-11e8-9499-8fcde80fba2c.png)  
Source: Hepta (Total number of cluster: 7)