nt_kmeans 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA256:
+  metadata.gz: 573007bfd7b624d0da1143cf1a95d2f8957cd1f26fdf7ffe82b8e762679a738b
+  data.tar.gz: f95b9d25c1a89be72a4ddf334bee83fa0bd1433d359c317165be9f9c6d623473
+SHA512:
+  metadata.gz: 0b58ed7e34ef55b61c657b644d38c698b7d99d6609ed3833da0c8b25908984bacfab902dae707830c70d65eac9d628b0e22f21b8ef8d312b6da4e5cbdc6141f0
+  data.tar.gz: a25a87418aa7e0b7d0ac16eec10bc2d0b3d4199fa3cf5123231aa72ccd9d43b97eb3fb8a4e2c509c9bbe5c74574a47b979013ce8742f9b12dc8815e547748f1f

data/lib/kmeans.rb

@@ -0,0 +1,87 @@
+class Kmeans
+  require 'json'
+
+  def self.hi
+    puts "K-means clustering ruby!"
+  end
+
+  def self.run k, data
+    # random first centers
+    centers = data.sample(k)
+    y = []
+    (0..(k-1)).each do |i|
+      y[i] = []
+    end
+    while true
+      # save pre-loop groups
+      y_old = y
+      # grouping
+      y = group_data(data, centers)
+      # break while loop if groups are not changed
+      if y === y_old
+        break
+      end
+      # update centers
+      centers = []
+      y.each do |ct|
+        centers.push(getCenter ct)
+      end
+    end
+    return centers, y
+  end
+
+  def self.group_data data, centers
+    y = []
+    (0..(centers.length-1)).each do |i|
+      y[i] = []
+    end
+    list = []
+    data.each do |x|
+      centers.each do |center|
+        list.push(euclidean_distance(x, center))
+      end
+      min = argmin(list)
+      y[min].push(x)
+      list = []
+    end
+    return y
+  end
+
+  # cosine
+  def self.euclidean_distance p1, p2
+    sum_1 = 0
+    sum_2 = 0
+    sum_12 = 0
+    p1.each_with_index do |p, i|
+      sum_1 = sum_1 + p**2
+      sum_12 = sum_12 + (p1[i].to_f * p2[i].to_f)
+    end
+    p2.each_with_index do |p|
+      sum_2 = sum_2 + p**2
+    end
+    return 1 - sum_12 / (Math.sqrt(sum_1) * Math.sqrt(sum_2))
+  end
+
+  def self.getCenter ps
+    p = []
+    ps.each do |p1|
+      p1.each_with_index do |p2, i|
+        p[i] = p[i].to_f+ p2.to_f
+      end
+    end
+    p.each_with_index do |p3, i|
+      p[i] = p[i]/ps.count
+    end
+    return p
+  end
+
+  def self.argmin arr
+    i = 0
+    (1..(arr.length-1)).each do |index|
+      if arr[i] >= arr[index]
+        i = index
+      end
+    end
+    return i
+  end
+end

metadata

@@ -0,0 +1,44 @@
+--- !ruby/object:Gem::Specification
+name: nt_kmeans
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+platform: ruby
+authors:
+- Ngoc Tuan
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2019-10-30 00:00:00.000000000 Z
+dependencies: []
+description: K-means clustering ruby
+email: tuansaker1412@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/kmeans.rb
+homepage: https://rubygems.org/gems/nt_kmeans
+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: 2.7.8
+signing_key: 
+specification_version: 4
+summary: K-means clustering ruby
+test_files: []