agglomerative_clustering 0.0.2 → 0.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 6368eb116e76afdf46d5dd16ffeaed04a3d590e8
-  data.tar.gz: 39447a07d51c37d5743dc910bb71ff011bfcd6d2
+  metadata.gz: 03755a284a11c8225365f9b22df38da433e07994
+  data.tar.gz: 6996369a3b16c5541cd694a8315c2c40606b91e3
 SHA512:
-  metadata.gz: 198911398132736c1cf06117e48e4f0026c94271586e8f172bf72733828f8bf1b218e03659b148663c56b2f8a3ca793e45b335527e90bc51577663221ddb4202
-  data.tar.gz: 5b04529399622524c8d7b36cd8b387c7dae73c83700e0e63455091e6357b39745bb78df0d01e6df0f8e29f0d264f2df89b23a79d9b744f2df393ceb10062249a
+  metadata.gz: 1061fb7aae2bc9c6cee7ba054593331a6e9db4c5388d895f821aad46b6d05d86721f61c947ce5377095ff7dc3adf8f93494ce9a3697f44357521d5ecc97e8523
+  data.tar.gz: f481ee1c97df283e3a354ed447319fbf9041d5507ac677269288e9c7802eb743510e016a6a0fca9a5232eb0a98420a1c393e1a08c9a52a0dfac91fcc21d91fd7

data/cluster.rb

@@ -56,3 +56,7 @@ clusters.each_with_index do |cluster, index|
     end
   end
 end
+
+puts 'Silhouette Coefficient of First Cluster: '
+sc = AgglomerativeClustering::SilhouetteCoefficient.new(clusters[0])
+puts sc.measure(clusters)

data/lib/agglomerative_clustering/distance_matrix.rb

@@ -1,3 +1,4 @@
+require 'matrix'
 module AgglomerativeClustering
   class DistanceMatrix
 
@@ -6,9 +7,7 @@ module AgglomerativeClustering
     end
 
     def matrix
-      Matrix.build(matrix_array.size, matrix_array.first.size) do |row, column|
-        matrix_array[row][column]
-      end
+      @matrix ||= build_matrix
     end
 
     def print_matrix
@@ -18,20 +17,38 @@ module AgglomerativeClustering
     def remove_edge index
       matrix_array.delete_at(index)
       matrix_array.each { |row| row.delete_at(index) }
-      Matrix.rows(matrix_array)
     end
 
     def add_edge weights
       matrix_array.each_with_index { |row, index|  row << weights[index] }
       matrix_array << weights
-      Matrix.rows(matrix_array)
+      @matrix = build_matrix
+    end
+
+    def shortest_distance
+      min_dist = 1.0/0
+      indexes = []
+      matrix.each_with_index do |index, row, column|
+        distance = matrix[row, column]
+        if distance < min_dist && (row != column)
+          min_dist = distance
+          indexes = [row, column]
+        end
+      end
+      indexes
     end
-    
+
     private
 
     def matrix_array
       @matrix_array ||= []
     end
 
+    def build_matrix
+      Matrix.build(matrix_array.size, matrix_array.first.size) do |row, column|
+        matrix_array[row][column]
+      end
+    end
+
   end
 end

data/lib/agglomerative_clustering/set.rb

@@ -1,4 +1,3 @@
-require 'matrix'
 module AgglomerativeClustering
   class Set
     include EuclideanDistance
@@ -26,11 +25,27 @@ module AgglomerativeClustering
 
     def cluster total_clusters
       while clusters.size > total_clusters
-        merge_clusters(shortest_distance)
+        merge_clusters(distance_matrix.shortest_distance)
       end
       clusters
     end
 
+    def outliers
+      set_outliers.uniq
+    end
+
+    def find_outliers percentage_of_clusters, distance
+      distance_matrix.matrix.each_with_index do |index, row, column|
+        count_hash[row] ||= 0
+        count_hash[row] += 1 if distance_matrix.matrix[row, column] > distance
+        set_outliers << points[row] if count_hash[row]/(distance_matrix.matrix.row_count - 1) > percentage_of_clusters/100
+      end
+      points.reject! { |point| outliers.include?(point) }
+      outliers
+    end
+
+    private
+
     def merge_clusters indexes
       index1, index2 = indexes
       new_cluster = clusters[index1].merge(clusters[index2])
@@ -48,22 +63,6 @@ module AgglomerativeClustering
       distance_matrix
     end
 
-    def outliers
-      set_outliers.uniq
-    end
-
-    def find_outliers percentage_of_clusters, distance
-      distance_matrix.matrix.each_with_index do |index, row, column|
-        count_hash[row] ||= 0
-        count_hash[row] += 1 if distance_matrix.matrix[row, column] > distance
-        set_outliers << points[row] if count_hash[row]/(distance_matrix.matrix.row_count - 1) > percentage_of_clusters/100
-      end
-      points.reject! { |point| outliers.include?(point) }
-      outliers
-    end
-
-    private
-
     def add_cluster new_cluster
       clusters << new_cluster
       update_distance_matrix(clusters.size - 1)
@@ -75,19 +74,6 @@ module AgglomerativeClustering
       distance_matrix.remove_edge(index)
     end
 
-    def shortest_distance
-      min_cluster_dist = 1.0/0
-      indexes = []
-      distance_matrix.matrix.each_with_index do |index, row, column|
-        distance = distance_matrix.matrix[row, column]
-        if distance < min_cluster_dist && distance != 0
-          min_cluster_dist = distance
-          indexes = [row, column]
-        end
-      end
-      indexes
-    end
-
     def set_outliers
       @set_outliers ||= []
     end

data/lib/agglomerative_clustering/silhouette_coefficient.rb

@@ -0,0 +1,38 @@
+module AgglomerativeClustering
+  class SilhouetteCoefficient
+    include EuclideanDistance
+    attr_reader :main_cluster
+
+    def initialize main_cluster
+      @main_cluster = main_cluster
+    end
+
+    def measure clusters
+      silhouettes = []
+      average_distances = []
+        main_cluster.points.each do |point1|
+          a1 = calculate_a1(point1)
+          (clusters - [main_cluster]).each do |cluster|
+            distances = []
+            cluster.points.each do |point2|
+              distances << euclidean_distance(point1, point2)
+            end
+            average_distances << distances.inject(:+)/distances.size
+          end
+          b1 = average_distances.min
+          s1 = (b1 - a1)/[a1,b1].max
+          silhouettes << s1
+        end
+      silhouettes.inject(:+) / silhouettes.size
+    end
+
+    def calculate_a1 point
+      distances = []
+      main_cluster.points.each do |point1|
+        distances << euclidean_distance(point, point1)
+      end
+      distances.inject(:+)/(distances.size - 1)
+    end
+
+  end
+end

data/lib/agglomerative_clustering/version.rb

@@ -1,3 +1,3 @@
 module AgglomerativeClustering
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

data/lib/agglomerative_clustering.rb

@@ -9,3 +9,4 @@ require "agglomerative_clustering/linkage/center"
 require "agglomerative_clustering/point"
 require "agglomerative_clustering/cluster"
 require "agglomerative_clustering/set"
+require "agglomerative_clustering/silhouette_coefficient"

data/spec/lib/agglomerative_clustering/distance_matrix_spec.rb

@@ -5,7 +5,7 @@ describe AgglomerativeClustering::DistanceMatrix do
       matrix = AgglomerativeClustering::DistanceMatrix.new(Matrix.empty)
       matrix.add_edge([1,2])
       matrix.add_edge([2,2,3])
-      expect(matrix.remove_edge(0)).to eql(Matrix[[2,3]])
+      expect(matrix.remove_edge(0)).to eql([[2,3]])
     end
   end
 
@@ -15,4 +15,11 @@ describe AgglomerativeClustering::DistanceMatrix do
       expect(matrix.add_edge([4,5,6,7])).to eql(Matrix[[1,2,3,4],[4,5,6,7]])
     end
   end
+
+  context '#shortest_distance' do
+    it 'will return the indexes of the shortest distances' do
+      matrix = AgglomerativeClustering::DistanceMatrix.new(Matrix.rows([[0,2,3],[2,0,3]]))
+      expect(matrix.shortest_distance).to eql([0,1])
+    end
+  end
 end

data/spec/lib/agglomerative_clustering/set_spec.rb

@@ -35,12 +35,6 @@ describe AgglomerativeClustering::Set do
     end
   end
 
-  context '#merge_clusters' do
-    it 'will merge two clusters into one and update the distance matrix' do
-      expect(@set.merge_clusters([0,1]).points).to eql([@point1, @point2])
-    end
-  end
-
   context '#find_outliers' do
     it 'will return a list of outliers' do
       outlier1 = FactoryGirl.build(:point, x:100, y:200, z:300)

data/spec/lib/agglomerative_clustering/silhouette_coefficient_spec.rb

@@ -0,0 +1,24 @@
+describe AgglomerativeClustering::SilhouetteCoefficient do
+
+  context '#measure' do
+    it 'will return the average silhoutte coefficient of a cluster' do
+      p1 = FactoryGirl.build(:point, x:1, y:1, z:1)
+      p2 = FactoryGirl.build(:point, x:3, y:3, z:3)
+      p3 = FactoryGirl.build(:point, x:17, y:17, z:17)
+      p4 = FactoryGirl.build(:point, x:16, y:16, z:16)
+      p5 = FactoryGirl.build(:point, x:18, y:18, z:18)
+      p6 = FactoryGirl.build(:point, x:2, y:2, z:2)
+      cluster1 = AgglomerativeClustering::Cluster.new(p1)
+      cluster2 = AgglomerativeClustering::Cluster.new(p2)
+      cluster3 = AgglomerativeClustering::Cluster.new(p3)
+      cluster4 = AgglomerativeClustering::Cluster.new(p4)
+      cluster5 = AgglomerativeClustering::Cluster.new(p5)
+      cluster6 = AgglomerativeClustering::Cluster.new(p6)
+      cluster1.merge(cluster2).merge(cluster3)
+      cluster4.merge(cluster5).merge(cluster6)
+      sc = AgglomerativeClustering::SilhouetteCoefficient.new(cluster1)
+      clusters = [cluster1, cluster4]
+      expect(sc.measure(clusters).round(4)).to eql(-0.0893)
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: agglomerative_clustering
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
 platform: ruby
 authors:
 - Bryan Mulvihill
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-10-30 00:00:00.000000000 Z
+date: 2014-11-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -79,6 +79,7 @@ files:
 - lib/agglomerative_clustering/linkage/single.rb
 - lib/agglomerative_clustering/point.rb
 - lib/agglomerative_clustering/set.rb
+- lib/agglomerative_clustering/silhouette_coefficient.rb
 - lib/agglomerative_clustering/version.rb
 - spec/factories/lib/agglomerative_clustering/cluster.rb
 - spec/factories/lib/agglomerative_clustering/point.rb
@@ -91,6 +92,7 @@ files:
 - spec/lib/agglomerative_clustering/linkage/complete_spec.rb
 - spec/lib/agglomerative_clustering/linkage/single_spec.rb
 - spec/lib/agglomerative_clustering/set_spec.rb
+- spec/lib/agglomerative_clustering/silhouette_coefficient_spec.rb
 - spec/spec_helper.rb
 homepage: https://github.com/bmulvihill/agglomerative_clustering
 licenses:
@@ -128,4 +130,5 @@ test_files:
 - spec/lib/agglomerative_clustering/linkage/complete_spec.rb
 - spec/lib/agglomerative_clustering/linkage/single_spec.rb
 - spec/lib/agglomerative_clustering/set_spec.rb
+- spec/lib/agglomerative_clustering/silhouette_coefficient_spec.rb
 - spec/spec_helper.rb