NDKMeans 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: d4c4f0b227178a699448686ead2882ff579ecf74
+  data.tar.gz: 9438ccf1c1bbda0cf556671e779189cae0907150
+SHA512:
+  metadata.gz: 1bdf489a238cf029f9fc2c674ae819486789b5e9b9b91403ced30d16b4758703f6c8570a0508392a153ea8a15ee5745d5b83b56c1854571c19d2066fd610807f
+  data.tar.gz: 9ce04af21cc3d7ff910ace41b1f57e00b2222170288b1d0ff253a3fca7067fc49cc6502b3f7eb586229777cd081194c77f46dc772ba8e018ba0dfa061f561b9f

data/Rakefile

@@ -0,0 +1,8 @@
+require 'rake/testtask'
+
+Rake::TestTask.new do |t|
+  t.libs << 'test'
+end
+
+desc "Run tests"
+task :default => :test

data/bin/TestMultiDimension

@@ -0,0 +1,10 @@
+#! /bin/ruby
+require 'MultiDimensional KMeans'
+
+test = MultiDimensionalKMeans.new 3, 4
+points = [[0,0,0,0], [1,1,1,1], [2,2,2,2], [3,3,3,3], [4,4,4,4], [1,2,3,4], [4,3,2,1]]
+centroids, clusters = test.build points 
+for i in 0...centroids.size
+	puts "CENTROID: " + centroids[i].to_s
+	puts "Cluster: " + clusters[i].to_s
+end

data/lib/MultiDimensional KMeans.rb

@@ -0,0 +1,190 @@
+$LOAD_PATH.unshift(File.expand_path('../../lib', __FILE__))
+require "gnuplot"
+require "rubygems"
+require 'csv'
+require 'pp'
+
+class Point
+	attr_accessor :values
+	def initialize(values)
+		@values = values
+	end
+	def to_s
+		return @values.to_s
+	end
+	def distance(p)
+		dis = 0
+		for i in 0...@values.length
+			dis = dis + (p.values[i] - @values[i])**2
+		end
+		ret = Math.sqrt(dis)
+	end
+	def equals(p)
+		eq = true
+		for i in 0...@values.size
+			if p.values[i] != @values[i]
+				eq = false
+			end
+		end
+		return eq
+	end
+end
+
+class MultiDimensionalKMeans
+	attr_accessor :centroids, :old_centroids, :clusters, :dimension
+	def initialize(inputs, dimen)
+		@dimension = dimen
+		@centroids = Array.new
+		for i in 0...inputs
+			@centroids[i] = Point.new zeros()
+		end
+	end
+
+	def zeros()
+		arr = Array.new
+		for i in 0...@dimension
+			arr[i] = 0
+		end
+		return arr
+	end
+
+	def generateKmeansPoints(points)
+		retArr = Array.new points[0].size
+		for i in 0...retArr.size
+			retArr[i] = Array.new
+		end
+		for i in 0...points[0].size
+			for j in 0...points.size
+				retArr[i][j] = points[j][i]
+			end
+		end
+		return retArr
+	end
+
+	def build(points)
+		# Calcular centroides aleatorios
+		kmeansPoints = generateKmeansPoints(points)
+		# Modificar función initialize_random_centroids
+		@old_centroids = Array.new @centroids
+		initialize_random_centroids kmeansPoints
+		while stop_condition
+			@clusters = Array.new @centroids.length
+			for i in 0...@clusters.length
+					clusters[i] = Array.new
+			end
+			# Meter cada punto a su cluster mas cercano
+			for i in 0...points.length
+				clusters[eval(points[i])].push Point.new(points[i])
+				# clusters[eval(x[i], y[i], z[i])].push Point.new x[i], y[i], z[i]
+			end
+			@old_centroids = Array.new @centroids
+			update_centroids kmeansPoints
+		end
+		cen, clus = @centroids, @clusters
+	end
+
+	def randomizeSubarray(arr)
+		random = rand(arr.max - arr.min) + arr.min
+	end
+
+	def randomizeCentroid(points)
+		centroidArray = Array.new points.size
+		for i in 0...points.size
+			centroidArray[i] = randomizeSubarray points[i]
+		end
+		return centroidArray
+	end
+
+	def initialize_random_centroids(points)
+		@old_centroids = Array.new @centroids
+		for i in 0...@centroids.length
+			arr = randomizeCentroid(points)
+			@centroids[i] = Point.new arr
+		end
+	end
+
+	def eval(point)
+		p = Point.new point
+		distancias = Array.new @centroids.length
+		for i in 0...distancias.length
+			distancias[i] = p.distance @centroids[i]
+		end
+		min_dis = distancias.min
+		min_idx = distancias.find_index min_dis
+		ret = min_idx
+	end
+
+	# Modificando
+	def update_centroids(points)
+		@old_centroids = Array.new @centroids
+		reinitialize = false
+		for i in 0...@clusters.length
+			puts "Cluster " + (i+1).to_s + " tiene " + @clusters[i].length.to_s
+			if @clusters[i].length < 2
+				reinitialize = true
+			end
+		end
+		if reinitialize
+			initialize_random_centroids points
+		else
+			minimus = Array.new points.size
+			maximus = Array.new points.size
+			for i in 0...@centroids.length
+					for j in 0...minimus.size
+						minimus[j] = Float::INFINITY
+						maximus[j] = 0
+					end
+					# min_x = Float::INFINITY
+					# min_y = Float::INFINITY
+					# min_z = Float::INFINITY
+					# max_x = 0
+					# max_y = 0
+					# max_z = 0
+					for j in 0...@clusters[i].length
+						# Verificar cada cluster para actualizar minimos y maximos
+						for k in 0...minimus.size
+							if @clusters[i][j].values[k] < minimus[k]
+								minimus[k] = @clusters[i][j].values[k]
+							elsif @clusters[i][j].values[k] > maximus[k]
+								maximus[k] = @clusters[i][j].values[k]
+							end
+						end
+						# if clusters[i][j].x < min_x
+							# min_x = clusters[i][j].x
+						# elsif clusters[i][j].x > max_x
+							# max_x = clusters[i][j].x
+						# end
+						# if clusters[i][j].y < min_y
+							# min_y = clusters[i][j].y
+						# elsif clusters[i][j].y > max_y
+							# max_y = clusters[i][j].y
+						# end
+						# if clusters[i][j].z < min_z
+							# min_z = clusters[i][j].z
+						# elsif clusters[i][j].z > max_z
+							# max_z = clusters[i][j].z
+						# end
+					end
+					newPoint = Array.new minimus.size
+					for x in 0...newPoint.size
+						newPoint[x] = (maximus[x]-minimus[x])/2 + minimus[x]
+					end
+					@centroids[i] = Point.new newPoint
+					# @centroids[i] = Point.new (max_x-min_x)/2+min_x, (max_y-min_y)/2+min_y, (max_z-min_z)/2+min_z
+				end
+		end
+	end
+
+	def stop_condition
+		ending = false
+		for i in 0...@centroids.size
+			fin = false
+			if !@centroids[i].equals @old_centroids[i]
+				ending = true
+				break
+			end
+		end
+		ret = ending
+	end
+
+end

metadata

@@ -0,0 +1,46 @@
+--- !ruby/object:Gem::Specification
+name: NDKMeans
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+platform: ruby
+authors:
+- Salvador Guerra Delgado
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2018-10-12 00:00:00.000000000 Z
+dependencies: []
+description: Multi dimensional K-Means clustering algorithm
+email: salvador.guerra.delgado2@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- Rakefile
+- bin/TestMultiDimension
+- lib/MultiDimensional KMeans.rb
+homepage: http://rubygems.org/gems/SEATC
+licenses:
+- MIT
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.5.2.1
+signing_key: 
+specification_version: 3
+summary: SEATC
+test_files: []