https://github.com/seregpie/nearestneighborchain

Builds a hierarchy of the clusters.
https://github.com/seregpie/nearestneighborchain

algorithm array cluster clustering collection distance hierarchical hierarchy javascript method object technique

Last synced: 20 days ago
JSON representation

Builds a hierarchy of the clusters.

• Host: GitHub
• URL: https://github.com/seregpie/nearestneighborchain
• Owner: SeregPie
• License: mit
• Created: 2016-12-02T13:47:56.000Z (about 8 years ago)
• Default Branch: master
• Last Pushed: 2020-02-28T14:39:51.000Z (almost 5 years ago)
• Last Synced: 2024-09-25T09:30:26.292Z (3 months ago)
• Topics: algorithm, array, cluster, clustering, collection, distance, hierarchical, hierarchy, javascript, method, object, technique
• Language: JavaScript
• Homepage:
• Size: 5.86 KB
• Stars: 3
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# NearestNeighborChain

`NearestNeighborChain(values, distance)`

Builds a hierarchy of the clusters.

| argument | description |

| ---: | :--- |

| `values` | An iterable of the values to build the hierarchy of the clusters from. |

| `distance` | A function to calculate the distance between two values. The value pairs with the lowest distance build a cluster. |

Returns the clustered values as a nested array.

## dependencies

- [BronKerbosch](https://github.com/SeregPie/BronKerbosch)

## setup

### npm

```shell

npm install @seregpie/nearest-neighbor-chain

```

### ES module

```javascript

import NearestNeighborChain from '@seregpie/nearest-neighbor-chain';

```

### Node

```javascript

let NearestNeighborChain = require('@seregpie/nearest-neighbor-chain');

```

### browser

```html

```

The module is globally available as `NearestNeighborChain`.

## usage

```javascript

let array = [4, 90, 12, 61, 29];

let clusters = NearestNeighborChain(array, (a, b) => Math.abs(a - b));

// => [[29, [4, 12]], [90, 61]]

```

---

Overlapping clusters are merged together.

```javascript

let intersection = function(a, b) {

  a = new Set(a);

  b = new Set(b);

  return [...a].filter(v => b.has(v));

};

let array = ['ac', 'ab', 'baab', 'aba', 'bc'];

let clusters = NearestNeighborChain(array, (a, b) => -intersection(a, b).length);

// => ['ac', 'bc', ['ab', 'baab', 'aba']]

```

## see also

- [KMeans](https://github.com/SeregPie/KMeans)