Program implementing the approximate version of DBSCAN introduced by Gan and Tao

Ivano Donadi

Last update: May 18, 2021

Overview

appr_dbscan_rust

Rust implementation of the approximate version of DBSCAN introduced by Gan and Tao in this paper

Notice

An upated version of this library is available in the linfa ecosistem

Recognized data files

Accepted data files should contain one data point per line and nothing else. Each line should contain the components of the point separated by whitespace.

Each component of a point will be read and stored as a 64 bit floating point variable

Example

1.0 1.1 0.5
2.3 3.4 6.2
...

Usage

There are four main functions in this library that differ in the kind of input they accept:

Warning

Each function below expects all points in the data file to have the same dimensionality and panics otherwise.

Approximated DBSCAN from data file with fixed dimensionality

If the dimensionality of each point is statically known (so not a result from another calculation) then this function can be used:

pub fn do_appr_dbscan_file<P, const D: usize>(
    filename: P, 
    epsilon: f64, 
    rho: f64, 
    min_pts: usize
) -> DBSCANResult<D> 
where
    P: AsRef<Path>,

Example

extern crate appr_dbscan;
use appr_dbscan::do_appr_dbscan_file;
use appr_dbscan::utils::DBSCANResult;
 
let res : DBSCANResult<2> = do_appr_dbscan_file("./datasets/out_test_1.txt", 0.3, 0.1, 10);
let clusters_count = res.len() - 1;
let noise_points_count = res[0].len();

Approximated DBSCAN from data file with unknown dimensionality

If the dimensionality of the data points is not statically known (like if there is a loop going through multiple files with different dimensionalities) then this function can be used:

pub fn do_appr_dbscan_auto_dimensionality_file<P>(
    filename: P, 
    epsilon: f64, 
    rho: f64, 
    min_pts: usize
) -> (VectorDBSCANResult, usize) 
where
    P: AsRef<Path>,

Example

extern crate appr_dbscan;
use appr_dbscan::do_appr_dbscan_auto_dimensionality_file;

let (res,dimensionality) = do_appr_dbscan_auto_dimensionality_file("./datasets/out_test_1.txt", 0.3, 0.1, 10);
println!("Points dimensionality: {}",dimensionality);
let clusters_count = res.len() - 1;
let noise_points_count = res[0].len();

Approximated DBSCAN from vector of points of fixed dimensionality

If you have a vector of points of the type Vec<[f64;D]> then this function can be used:

pub fn do_appr_dbscan_points<const D: usize>(
    points: Vec<Point<D>>, 
    epsilon: f64, 
    rho: f64, 
    min_pts: usize
) -> DBSCANResult<D>

Example

extern crate appr_dbscan;
use appr_dbscan::do_appr_dbscan_points;
use appr_dbscan::utils::DBSCANResult;
 
let points = vec![[0.0,0.0],[1.0,1.0],[0.0,1.0],[1.0,0.0],[2.0,1.0],[0.0,2.0],[2.0,1.0],[1.0,1.0]];
let res : DBSCANResult<2> = do_appr_dbscan_points(points, 0.3, 0.1, 10);
let clusters_count = res.len() - 1;
let noise_points_count = res[0].len();

Approximated DBSCAN from vector of points of unknown dimensionality

If you have a vector of points of the type Vec<Vec<f64>> (in example if you are in a loop clustering different vectors) then this function can be used:

pub fn do_appr_dbscan_auto_dimensionality_points(
    points: Vec<VectorPoint>, 
    epsilon: f64, 
    rho: f64, 
    min_pts: usize
) -> (VectorDBSCANResult, usize)

Example

extern crate appr_dbscan;
use appr_dbscan::do_appr_dbscan_auto_dimensionality_points;

let points = vec![vec![0.0,0.0],vec![1.0,1.0],vec![0.0,1.0],vec![1.0,0.0],vec![2.0,1.0],vec![0.0,2.0],vec![2.0,1.0],vec![1.0,1.0]];
let (res, dimensionality) = do_appr_dbscan_auto_dimensionality_points(points, 0.3, 0.1, 10);
println!("Points dimensionality: {}",dimensionality);
let clusters_count = res.len() - 1;
let noise_points_count = res[0].len();

You might also like...

A library implementing GPU-accelerated cryptographic functionality for the zkSync prover.

Program implementing the approximate version of DBSCAN introduced by Gan and Tao

Related tags

Overview

appr_dbscan_rust

Notice

Recognized data files

Example

Usage

Warning

Approximated DBSCAN from data file with fixed dimensionality

Example

Approximated DBSCAN from data file with unknown dimensionality

Example

Approximated DBSCAN from vector of points of fixed dimensionality

Example

Approximated DBSCAN from vector of points of unknown dimensionality

Example

You might also like...

A library implementing GPU-accelerated cryptographic functionality for the zkSync prover.

A small, basical and unoptimized version of RWKV in Rust written by someone with no math or ML knowledge.

Small program which groups images based on the GPS position.

Narwhal and Tusk A DAG-based Mempool and Efficient BFT Consensus.

MesaTEE GBDT-RS : a fast and secure GBDT library, supporting TEEs such as Intel SGX and ARM TrustZone

Ecosystem of libraries and tools for writing and executing extremely fast GPU code fully in Rust.

Ecosystem of libraries and tools for writing and executing fast GPU code fully in Rust.

[WIP] An experimental Java-like language and it's virtual machine, for learning Java and JVM.

Robust and Fast tokenizations alignment library for Rust and Python

Owner

Ivano Donadi

A naive DBSCAN implementation in Rust

A pure, low-level tensor program representation enabling tensor program optimization via program rewriting

HNSW ANN from the paper "Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphs"

Graph-based Approximate Nearest Neighbor Search

🚀 efficient approximate nearest neighbor search algorithm collections library written in Rust 🦀 .

Rust-port of spotify/annoy as a wrapper for Approximate Nearest Neighbors in C++/Python optimized for memory usage.

A neural network model that can approximate any non-linear function by using the random search algorithm for the optimization of the loss function.

Rust-port of spotify/annoy as a wrapper for Approximate Nearest Neighbors in C++/Python optimized for memory usage.

An approximate K-NN written in Rust.

A toolkit for implementing polynomial IOPs (PIOPs)