generate test data for K-means clustering

kmeans_sample_data.rb

require 'pp'

unless ARGV.size == 2
  $stderr.puts "Usage: ruby #{__FILE__} <n> <expected_k>"
  raise "invalid arguments"
end

VARIANCE = 0.05

def draw_gaussian
  r1 = rand
  r2 = rand
  x = Math.sqrt( -2 * Math.log(r1) ) * Math.cos(2*Math::PI*r2)
  y = Math.sqrt( -2 * Math.log(r1) ) * Math.sin(2*Math::PI*r2)
  [x,y]
end

def draw_random_from_gaussian(n,k)
  center = Array.new(k) {|i| [rand, rand] }

  points = Array.new(n) do |i|
    cx, cy = center[ rand(k) ]
    dx, dy = draw_gaussian.map {|x| x * VARIANCE }
    [cx+dx, cy+dy].map {|x| x - x.floor }
  end
  points
end

points = draw_random_from_gaussian( ARGV[0].to_i , ARGV[1].to_i)
$stdout.puts points.map {|point| point.join(' ') }