kmeans-clusterer 0.10.0 → 0.11.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 873f13d79f2d199400d1e359aca6e56c74d68ac0
-  data.tar.gz: a4d1f06a71b2e8289c60badeb93fb007fee06af6
+  metadata.gz: 2c2552e6a8ee7eddd6d03d8fa8cf7038c0458a11
+  data.tar.gz: 62172e9d841aa4df7b332a4e2cda55bf426bc9b1
 SHA512:
-  metadata.gz: 1e160f4bbe512e7aac37c2edcbcc63c1f103e51ae4c441a0c2f2e6cc57783344eddbcd31f30ea1a9a32dcc979250fb5bab6e1faba7a1d686612362d4aa798129
-  data.tar.gz: 405a6ab16edcb0fcdd958c9963446a44bd2954be9cdb1e95dc728bf7d10e18536334c693ca38f00b956d51358b3ab105aa2a9a3f68c862ab99914e4e4ee894e4
+  metadata.gz: a33d17f4749b00e7ee8a0829e6a402db6eb768bd62f5dcb1e3ed9dcf3020b5ec074b3e5e38a817fb6d4a51ec13b8cf75f189a55b6a3e488738b4dc2828cbb878
+  data.tar.gz: c4a2e27da37141435583d1b09cbf4001f4b2c44a5b2900155e2444f45d9663edf0ccab6d4f80c50d05ad85ce31710a7d4558356bded561bf17199e7333c3e14d

data/lib/kmeans-clusterer.rb

@@ -46,12 +46,13 @@ class KMeansClusterer
 
 
   class Point
-    attr_reader :id, :data
+    attr_reader :id, :data, :centroid_distances
     attr_accessor :cluster, :label
 
-    def initialize id, data, label = nil
+    def initialize id, data, centroid_distances, label = nil
       @id = id
       @data = data
+      @centroid_distances = centroid_distances
       @label = label
     end
 
@@ -70,6 +71,10 @@ class KMeansClusterer
     def dimension
       @data.length
     end
+
+    def centroid_distance
+      @centroid_distances[@cluster.id]
+    end
   end
 
 
@@ -116,7 +121,7 @@ class KMeansClusterer
       opts[:std] = std
     end
 
-    opts[:points_matrix] = data
+    opts[:data] = data
     opts[:row_norms] = Scaler.row_norms(data)
 
     bestrun = nil
@@ -137,7 +142,7 @@ class KMeansClusterer
   end
 
 
-  attr_reader :k, :points, :clusters, :centroids, :error, :mean, :std, :iterations, :runtime
+  attr_reader :k, :points, :clusters, :centroids, :error, :mean, :std, :iterations, :runtime, :distances, :data
 
 
   def initialize opts = {}
@@ -146,8 +151,8 @@ class KMeansClusterer
     @labels = opts[:labels] || []
     @row_norms = opts[:row_norms]
 
-    @points_matrix = opts[:points_matrix]
-    @points_count = @points_matrix.shape[1] if @points_matrix
+    @data = opts[:data]
+    @points_count = @data.shape[1] if @data
     @mean = opts[:mean]
     @std = opts[:std]
     @scale_data = opts[:scale_data]
@@ -167,10 +172,10 @@ class KMeansClusterer
       @iterations +=1
 
       min_distances.fill! Float::INFINITY
-      distances = Distance.euclidean(@centroids, @points_matrix, @row_norms)
+      @distances = Distance.euclidean(@centroids, @data, @row_norms)
 
       @k.times do |cluster_id|
-        dist = NArray.ref distances[true, cluster_id].flatten
+        dist = NArray.ref @distances[true, cluster_id].flatten
         mask = dist < min_distances
         @cluster_assigns[mask] = cluster_id
         min_distances[mask] = dist[mask]
@@ -183,7 +188,7 @@ class KMeansClusterer
         point_ids = @cluster_assigns.eq(cluster_id).where
 
         unless point_ids.empty?
-          points = @points_matrix[true, point_ids]
+          points = @data[true, point_ids]
           newcenter = points.mean(1)
           move = Distance.euclidean(centroid, newcenter)
           max_move = move if move > max_move
@@ -203,17 +208,21 @@ class KMeansClusterer
   def finish
     @clusters = @k.times.map do |i|
       centroid = NArray.ref @centroids[true, i].flatten
-      Cluster.new i, Point.new(-i, centroid)
+      Cluster.new i, Point.new(-1, centroid, nil, nil)
     end
 
     @points = @points_count.times.map do |i|
-      data = NArray.ref @points_matrix[true, i].flatten
-      point = Point.new(i, data, @labels[i])
+      data = NArray.ref @data[true, i].flatten
+      point = Point.new(i, data, @distances[i, true], @labels[i])
       cluster = @clusters[@cluster_assigns[i]]
-      cluster.points << point
+      cluster << point
       point
     end
 
+    @clusters.each do |c| 
+      c.points.sort_by! &:centroid_distance
+    end
+
     self
   end
 
@@ -236,17 +245,13 @@ class KMeansClusterer
   def silhouette
     return 1.0 if @k < 2
 
-    distances = Distance.euclidean(@centroids, @points_matrix, @row_norms)
+    scores = @points.map do |point|
+      sort_index = point.centroid_distances.sort_index
+      c1_points = get_points_for_cluster sort_index[0]
+      c2_points = get_points_for_cluster sort_index[1]
 
-    scores = @points_count.times.map do |i|
-      point = get_point i
-      cluster_indexes = distances[i, true].sort_index
-
-      c1_points = get_points_for_centroid cluster_indexes[0]
-      c2_points = get_points_for_centroid cluster_indexes[1]
-
-      a = dissimilarity(c1_points, point)
-      b = dissimilarity(c2_points, point)
+      a = dissimilarity(c1_points, point.data)
+      b = dissimilarity(c2_points, point.data)
       (b - a) / [a,b].max
     end
 
@@ -282,9 +287,9 @@ class KMeansClusterer
       centroid_ids << pick
 
       while centroid_ids.length < @k
-        centroids = @points_matrix[true, centroid_ids]
+        centroids = @data[true, centroid_ids]
 
-        distances = Distance.euclidean(centroids, @points_matrix, @row_norms)
+        distances = Distance.euclidean(centroids, @data, @row_norms)
 
         d2 = []
         @points_count.times do |i|
@@ -300,7 +305,7 @@ class KMeansClusterer
         centroid_ids << pick
       end
 
-      @centroids = @points_matrix[true, centroid_ids]
+      @centroids = @data[true, centroid_ids]
     end
 
     def custom_centroid_init
@@ -309,24 +314,20 @@ class KMeansClusterer
     end
 
     def random_centroid_init
-      @centroids = @points_matrix[true, pick_k_random_indexes]
+      @centroids = @data[true, pick_k_random_indexes]
     end
 
     def pick_k_random_indexes
       @points_count.times.to_a.sample @k
     end
 
-    def get_point i
-      NArray.ref @points_matrix[true, i].flatten
-    end
-
     def get_centroid i
       NArray.ref(@centroids[true, i].flatten)
     end
 
-    def get_points_for_centroid i
+    def get_points_for_cluster i
       point_ids = @cluster_assigns.eq(i).where
-      points = @points_matrix[true, point_ids]
+      points = @data[true, point_ids]
       points.empty? ? NArray.sfloat(0) : NArray.ref(points)
     end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: kmeans-clusterer
 version: !ruby/object:Gem::Version
-  version: 0.10.0
+  version: 0.11.0
 platform: ruby
 authors:
 - Geoff Buesing
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-03-10 00:00:00.000000000 Z
+date: 2015-03-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: narray