K-Means

example1	example2

https://yja938882.github.io/d3_practice/k_means.html

Load

<script type="text/javascript" src="/* path */k_means.js"></script>

DataFormat

var Data = [
 [0.5, 0.7],
 [3.6, 0.11],
 ...
];

Docs

  Constructor

• K_Means(args)

var kmeans = K_Means( { k : 3, data : [[0.5], [0.6], ... ] });

• args Object
  • k Number : # of clusters.
  • data Array : Array of data points.

   Methods

• clustering(dist, callback1, callback2, callback3)

/* Clustring. */
kmeans.clustering(kmeans.euclidean, render1, render2, render3);

• dist Function : Distance function.
• callback1 Function : Function that is called after Assignment Step.
• callback2 Function : Function that is called after Update Step.
• callback3 Function : Function that is called after clustering.

• setNearestCluster(dist, callback)

/* Assignment Step. Assign each observation to the cluster. */
kmeans.setNearsetCluster(kmeans.euclidean, render);

• dist Function : Distance function.
• callback Function : Function that is called after Update Step.

• updateCentroid(callback)

/* Update Step. Calculate the new means to be the centroids */
kmeans.updateCentroid(render);

• callback Function : Function that is called after Update Step.

• euclidean(pointA, pointB)

/* Euclidean distance function. */
var dist = kmeans.euclidean(pointA, pointB);

• pointA Array<Number> : point.
• pointB Array<Number> : point.

• manhattan(pointA, pointB)

/* Manhattan distance function. */
var dist = kmeans.manhattan( pointA, pointB);

• pointA Array<Number> : point.
• pointB Array<Number> : point.