ml-hclust

hclust

Hierarchical clustering algorithms in JavaScript

Installation

npm install ml-hclust

Methods

Generate a clustering hierarchy.

• AGNES (AGglomerative NESting): Continuously merge nodes that have the least dissimilarity.
• DIANA (Divisive ANAlysis): The process starts at the root with all the points as one cluster and recursively splits the higher level clusters to build the dendrogram.
• BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies): Incrementally construct a CF (Clustering Feature) tree, a hierarchical data structure for multiphase clustering
• CURE (Clustering Using REpresentatives):
• CHAMELEON

new Hclust([options])

Creates a new Hclust instance with the given parameters or the default ones.

Arguments

• options - Object with options for the algorithm

Options

• name: The name of the hierarchical cluster, it could be agnes(default), diana, birch, cure or chameleon.
• sim: The kind of distance or similarity to use between vectors, the default is euclidean, but for a complete list see ml-distance.
• kind: The kind of similarity to use between clusters, it could be single(default), complete, average, centroid or ward

Example

var HCL = require('ml-hclust');

// actually this are the default values
var options = {
    name: 'agnes',
    sim: 'euclidean',
    kind: 'single'
};

var agnes = new HCL(options);

cluster(data)

Creates the tree based in the data points.

Arguments

• data: Array of points to be clustered.

Example

var data = [[2,6], [3,4], [3,8]];
var myHcl = new HCL();
myHcl.cluster(data);

getLevel(L)

Returns L level of the hierarchical clustering tree.

Arguments

• L - Level of the hierarchical tree.

Example

// the clustering points
var agnesData = [[2,6], [3,4], [3,8], [4,5], [4,7], [6,2], [7,2], [7,4], [8,4], [8,5]];

// creates the hierarchical tree
var HC = new Hclust();
HC.cluster(agnesData);

// the array of clusters
var ansAgnes = HC.getLevel(3);
/*
[ [ [ 3, 4 ], [ 4, 5 ], [ 3, 8 ], [ 4, 7 ] ],
  [ [ 6, 2 ], [ 7, 2 ] ],
  [ [ 7, 4 ], [ 8, 4 ], [ 8, 5 ] ],
  [ [ 2, 6 ] ] ]
*/

export()

Exports the model to a JSON object that can be written to disk and reloaded

load(model)

Returns a new Hclust instance based on the model.

Arguments

• model - JSON object generated with Hclust.export()

Test

$ npm install
$ npm test

Authors

• Miguel Asencio

License

MIT