Last active
May 27, 2019 05:36
-
-
Save jbeuckm/5731501 to your computer and use it in GitHub Desktop.
K-Means in 2D
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
<html lang="en" xmlns:m="http://www.w3.org/1998/Math/MathML"> | |
<head> | |
<meta charset="utf-8"> | |
<link rel="stylesheet" href="https://rawgithub.com/yannickulrich/presenter/master/lib/jqMath/UnifrakturMaguntia.css"> | |
<link rel="stylesheet" href="https://rawgithub.com/yannickulrich/presenter/master/lib/jqMath/jqmath-0.4.0.css"> | |
<script src="http://code.jquery.com/jquery-1.4.3.min.js"></script> | |
<script src="https://rawgithub.com/yannickulrich/presenter/master/lib/jqMath/jqmath-etc-0.4.0.min.js"></script> | |
<style> | |
div { | |
position: absolute; | |
} | |
line { | |
stroke-width: .4px; | |
} | |
#stats-table { | |
width: 460px; | |
border: none; | |
left: 0px; | |
} | |
</style> | |
</head> | |
<body> | |
<div id="graph"></div> | |
<div id="controls"> | |
<button onclick="step();">Step</button> | |
<button onclick="changeK(1);">Increase K</button> | |
<button onclick="changeK(-1);">Decrease K</button> | |
<table id="stats-table"> | |
<tr> | |
<td> | |
$∣ \{ x_i : {{x_i}^{t-1}∈c_k , {x_i}^{t}∉c_k } \} ∣$ | |
</td> | |
<td> | |
$∑↙{k=1}↖K ∑↙{x_i∈c_k} {‖ x_i - μ_k ‖^2}$ | |
</td> | |
<!-- | |
<td> | |
$\ov{(m_i-x)}∣x∈S_i$ | |
</td> | |
<td> | |
$\ov{(m_i-m_j)}$ | |
</td> | |
--> | |
</tr> | |
</table> | |
</div> | |
<script src="unpkg.com/@jbeuckm/k-means-js"></script> | |
<script type="text/javascript" src="http://mbostock.github.com/d3/d3.js"></script> | |
<script> | |
var ranges = [ [0,10], [0,10] ]; | |
var points = kmeans.generateRandomPoints(ranges, 1000); | |
var means = []; | |
var assignments = kmeans.assignPointsToMeans(points, means); | |
var svg = d3.select('#graph').append('svg').attr('width',960).attr('height',500); | |
var graph = svg.append('g').attr('transform', 'translate(460,0)'); | |
var meanLayer = graph.append('g'); | |
var xScale = d3.scale.linear().domain([0,10]).range([0,500]); | |
var yScale = d3.scale.linear().domain([0,10]).range([0,500]); | |
var color = d3.scale.category10(); | |
function redraw() { | |
var pointDots = graph.selectAll('.pointDots').data(points); | |
pointDots.enter().append('circle').attr('class','pointDots') | |
.attr('r', 2) | |
.attr('cx',function(d){ return xScale(d[0]); }) | |
.attr('cy',function(d){ return yScale(d[1]); }); | |
var assignmentLines = meanLayer.selectAll('.assignmentLines').data(assignments); | |
assignmentLines.enter().append('line').attr('class','assignmentLines') | |
.attr('x1',function(d, i){ return xScale(points[i][0]); }) | |
.attr('y1',function(d, i){ return yScale(points[i][1]); }) | |
.attr('x2',function(d, i){ return xScale(means[d][0]); }) | |
.attr('y2',function(d, i){ return yScale(means[d][1]); }) | |
.attr('stroke', function(d) { return color(d); }); | |
assignmentLines.transition().duration(500) | |
.attr('x2',function(d, i){ return xScale(means[d][0]); }) | |
.attr('y2',function(d, i){ return yScale(means[d][1]); }) | |
.attr('stroke', function(d) { return color(d); }); | |
var meanDots = meanLayer.selectAll('.meanDots').data(means); | |
meanDots.enter().append('circle').attr('class','meanDots') | |
.attr('r', 5) | |
.attr('stroke', function(d, i) { return color(i); }) | |
.attr('stroke-width', 3) | |
.attr('fill', 'white') | |
.attr('cx',function(d){ return xScale(d[0]); }) | |
.attr('cy',function(d){ return yScale(d[1]); }); | |
meanDots.transition().duration(500) | |
.attr('cx',function(d){ return xScale(d[0]); }) | |
.attr('cy',function(d){ return yScale(d[1]); }); | |
meanDots.exit().remove(); | |
} | |
changeK(5); | |
redraw(); | |
function step() { | |
oldAssignments = assignments; | |
kmeans.moveMeansToCenters(points, assignments, means); | |
assignments = kmeans.assignPointsToMeans(points, means); | |
var changeCount = kmeans.countChangedAssignments(assignments, oldAssignments); | |
var aveDistance = kmeans.findAverageDistancePointToMean(points, means, assignments); | |
var aveMeanSeparation = kmeans.findAverageMeanSeparation(means); | |
var sse = kmeans.sumSquaredError(points, means, assignments); | |
var row = d3.select('#stats-table').append('tr'); | |
row.append('td').html(changeCount); | |
row.append('td').html(sse); | |
redraw(); | |
} | |
function changeK(amt) { | |
if (amt > 0) { | |
while (amt--) { | |
var i = Math.floor(Math.random() * points.length); | |
var p = points[i]; | |
var newPoint = p.slice(0); | |
console.log("adding new point "+newPoint); | |
means.push(newPoint); | |
} | |
} | |
else while (amt < 0) { | |
means.pop(); | |
amt++; | |
} | |
assignments = kmeans.assignPointsToMeans(points, means); | |
redraw(); | |
} | |
</script> | |
</body> | |
</html> |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment