nirvdrum-ai4r 1.9.1

data/test/data/data_set.csv

@@ -0,0 +1,121 @@
+zone,rooms,size,price
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[36K-45K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[36K-45K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[36K-45K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[29K-35K]
+Moron Sur (GBA),2,[28 m2 - 39 m2],[36K-45K]
+Moron Sur (GBA),2,[40 m2 - 50 m2],[36K-45K]
+Moron Sur (GBA),2,[40 m2 - 50 m2],[29K-35K]
+Moron Sur (GBA),2,[40 m2 - 50 m2],[29K-35K]
+Moron Sur (GBA),2,[40 m2 - 50 m2],[56K-75K]
+Moron Sur (GBA),3,[40 m2 - 50 m2],[36K-45K]
+Moron Sur (GBA),2,[40 m2 - 50 m2],[29K-35K]
+Moron Sur (GBA),2,[51 m2 - 59 m2],[36K-45K]
+Moron Sur (GBA),3,[51 m2 - 59 m2],[46K-55K]
+Moron Sur (GBA),3,[51 m2 - 59 m2],[36K-45K]
+Moron Sur (GBA),3,[51 m2 - 59 m2],[36K-45K]
+Moron Sur (GBA),3,[51 m2 - 59 m2],[36K-45K]
+Moron Sur (GBA),3,[51 m2 - 59 m2],[36K-45K]
+Moron Sur (GBA),3,[51 m2 - 59 m2],[46K-55K]
+Moron Sur (GBA),2,[51 m2 - 59 m2],[46K-55K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[46K-55K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[46K-55K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[56K-75K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[56K-75K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[36K-45K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[36K-45K]
+Moron Sur (GBA),3,[60 m2 - 70 m2],[56K-75K]
+Moron Sur (GBA),3,[71 m2 - 85 m2],[46K-55K]
+Moron Sur (GBA),3,[86 m2 - 100 m2],[46K-55K]
+Moron Sur (GBA),4,[101 m2 - 125 m2],[101K-125K]
+Moron Sur (GBA),3,[101 m2 - 125 m2],[76K-100K]
+Moron Sur (GBA),4,[101 m2 - 125 m2],[170K-200K]
+Moron Sur (GBA),4,[101 m2 - 125 m2],[76K-100K]
+Moron Sur (GBA),4,[126 m2 - 160 m2],[56K-75K]
+Moron Sur (GBA),4,[126 m2 - 160 m2],[126K-150K]
+Recoleta (CABA),2,[28 m2 - 39 m2],[46K-55K]
+Recoleta (CABA),2,[28 m2 - 39 m2],[56K-75K]
+Recoleta (CABA),2,[28 m2 - 39 m2],[56K-75K]
+Recoleta (CABA),2,[28 m2 - 39 m2],[46K-55K]
+Recoleta (CABA),2,[40 m2 - 50 m2],[56K-75K]
+Recoleta (CABA),2,[40 m2 - 50 m2],[76K-100K]
+Recoleta (CABA),2,[40 m2 - 50 m2],[76K-100K]
+Recoleta (CABA),2,[40 m2 - 50 m2],[56K-75K]
+Recoleta (CABA),2,[40 m2 - 50 m2],[101K-125K]
+Recoleta (CABA),2,[51 m2 - 59 m2],[101K-125K]
+Recoleta (CABA),2,[51 m2 - 59 m2],[56K-75K]
+Recoleta (CABA),2,[51 m2 - 59 m2],[76K-100K]
+Recoleta (CABA),2,[51 m2 - 59 m2],[76K-100K]
+Recoleta (CABA),2,[60 m2 - 70 m2],[56K-75K]
+Recoleta (CABA),2,[60 m2 - 70 m2],[76K-100K]
+Recoleta (CABA),3,[60 m2 - 70 m2],[101K-125K]
+Recoleta (CABA),2,[60 m2 - 70 m2],[76K-100K]
+Recoleta (CABA),2,[60 m2 - 70 m2],[76K-100K]
+Recoleta (CABA),2,[60 m2 - 70 m2],[126K-150K]
+Recoleta (CABA),3,[60 m2 - 70 m2],[76K-100K]
+Recoleta (CABA),3,[60 m2 - 70 m2],[76K-100K]
+Recoleta (CABA),2,[71 m2 - 85 m2],[152K-169K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[126K-150K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[152K-169K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[170K-200K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[101K-125K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[170K-200K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[76K-100K]
+Recoleta (CABA),2,[71 m2 - 85 m2],[126K-150K]
+Recoleta (CABA),3,[71 m2 - 85 m2],[126K-150K]
+Recoleta (CABA),2,[71 m2 - 85 m2],[200K-275K]
+Recoleta (CABA),3,[86 m2 - 100 m2],[152K-169K]
+Recoleta (CABA),3,[86 m2 - 100 m2],[170K-200K]
+Recoleta (CABA),4,[86 m2 - 100 m2],[126K-150K]
+Recoleta (CABA),4,[86 m2 - 100 m2],[126K-150K]
+Recoleta (CABA),3,[86 m2 - 100 m2],[200K-275K]
+Recoleta (CABA),3,[86 m2 - 100 m2],[126K-150K]
+Recoleta (CABA),3,[86 m2 - 100 m2],[152K-169K]
+Recoleta (CABA),3,[86 m2 - 100 m2],[200K-275K]
+Recoleta (CABA),3,[101 m2 - 125 m2],[200K-275K]
+Recoleta (CABA),3,[101 m2 - 125 m2],[170K-200K]
+Recoleta (CABA),4,[101 m2 - 125 m2],[170K-200K]
+Recoleta (CABA),4,[101 m2 - 125 m2],[200K-275K]
+Recoleta (CABA),3,[101 m2 - 125 m2],[56K-75K]
+Recoleta (CABA),4,[101 m2 - 125 m2],[200K-275K]
+Recoleta (CABA),4,[101 m2 - 125 m2],[170K-200K]
+Recoleta (CABA),3,[126 m2 - 160 m2],[200K-275K]
+Recoleta (CABA),4,[126 m2 - 160 m2],[200K-275K]
+Recoleta (CABA),3,[126 m2 - 160 m2],[152K-169K]
+Recoleta (CABA),3,[126 m2 - 160 m2],[200K-275K]
+Recoleta (CABA),4,[126 m2 - 160 m2],>275K
+Recoleta (CABA),3,[126 m2 - 160 m2],[200K-275K]
+Recoleta (CABA),4,[126 m2 - 160 m2],>275K
+Recoleta (CABA),3,[126 m2 - 160 m2],[170K-200K]
+Tigre (GBA),2,[28 m2 - 39 m2],[56K-75K]
+Tigre (GBA),2,[40 m2 - 50 m2],[56K-75K]
+Tigre (GBA),2,[40 m2 - 50 m2],[76K-100K]
+Tigre (GBA),2,[51 m2 - 59 m2],[101K-125K]
+Tigre (GBA),2,[51 m2 - 59 m2],[101K-125K]
+Tigre (GBA),3,[51 m2 - 59 m2],[101K-125K]
+Tigre (GBA),3,[51 m2 - 59 m2],[101K-125K]
+Tigre (GBA),3,[51 m2 - 59 m2],[126K-150K]
+Tigre (GBA),2,[71 m2 - 85 m2],[126K-150K]
+Tigre (GBA),2,[71 m2 - 85 m2],[126K-150K]
+Tigre (GBA),2,[71 m2 - 85 m2],[126K-150K]
+Tigre (GBA),3,[86 m2 - 100 m2],[170K-200K]
+Tigre (GBA),4,[86 m2 - 100 m2],[152K-169K]
+Tigre (GBA),4,[86 m2 - 100 m2],[152K-169K]
+Tigre (GBA),3,[86 m2 - 100 m2],[152K-169K]
+Tigre (GBA),3,[86 m2 - 100 m2],[152K-169K]
+Tigre (GBA),3,[101 m2 - 125 m2],[152K-169K]
+Tigre (GBA),4,[101 m2 - 125 m2],[170K-200K]
+Tigre (GBA),4,[101 m2 - 125 m2],[170K-200K]
+Tigre (GBA),4,[101 m2 - 125 m2],[200K-275K]
+Tigre (GBA),4,[101 m2 - 125 m2],[200K-275K]
+Tigre (GBA),4,[101 m2 - 125 m2],[170K-200K]
+Tigre (GBA),4,[101 m2 - 125 m2],>275K
+Tigre (GBA),4,[101 m2 - 125 m2],>275K
+Tigre (GBA),4,[126 m2 - 160 m2],>275K

data/test/data/data_set_test.rb

@@ -0,0 +1,96 @@
+# Author::    Sergio Fierens
+# License::   MPL 1.1
+# Project::   ai4r
+# Url::       http://ai4r.rubyforge.org/
+#
+# You can redistribute it and/or modify it under the terms of 
+# the Mozilla Public License version 1.1  as published by the 
+# Mozilla Foundation at http://www.mozilla.org/MPL/MPL-1.1.txt
+
+require 'test/unit'
+require File.dirname(__FILE__) + '/../../lib/ai4r/data/data_set'
+
+module Ai4r
+  module Data
+    class DataSetTest < Test::Unit::TestCase
+      
+      def test_load_csv_with_labels
+        set = DataSet.new.load_csv_with_labels("#{File.dirname(__FILE__)}/data_set.csv")
+        assert_equal 120, set.data_items.length
+        assert_equal ["zone", "rooms", "size", "price"], set.data_labels
+      end
+      
+      def test_build_domains
+        domains =  [  Set.new(["New York", "Chicago"]), 
+                      Set.new(["M", "F"]), 
+                      [5, 85],
+                      Set.new(["Y", "N"]) ]
+        data = [  [ "New York", "M", 23, "Y"],
+                  [ "Chicago", "M", 85, "Y"],
+                  [ "New York", "F", 32, "Y"],
+                  [ "New York", "M", 5, "N"],
+                  [ "Chicago", "M", 15, "N"],
+                  [ "Chicago", "F", 45, "Y"] ]
+        labels = ["city", "gender", "age", "result"]
+        set = DataSet.new({:data_items => data, :data_labels => labels})
+        assert_equal domains, set.build_domains
+        assert_equal domains[0], set.build_domain("city")
+        assert_equal domains[1], set.build_domain(1)
+        assert_equal domains[2], set.build_domain("age")
+        assert_equal domains[3], set.build_domain("result")
+      end
+      
+      def test_set_data_labels
+        labels = ["A", "B"]
+        set = DataSet.new.set_data_labels(labels)
+        assert_equal labels, set.data_labels
+        set = DataSet.new(:data_labels => labels)
+        assert_equal labels, set.data_labels
+        set = DataSet.new(:data_items => [[ 1, 2, 3]])
+        assert_raise(ArgumentError) { set.set_data_labels(labels) }
+      end
+
+      def test_set_data_items
+        items = [  [ "New York", "M", "Y"],
+                  [ "Chicago", "M", "Y"],
+                  [ "New York", "F", "Y"],
+                  [ "New York", "M", "N"],
+                  [ "Chicago", "M", "N"],
+                  [ "Chicago", "F", "Y"] ]
+        set = DataSet.new.set_data_items(items)
+        assert_equal items, set.data_items
+        assert_equal 3, set.data_labels.length
+        items << items.first[0..-2]
+        assert_raise(ArgumentError) { set.set_data_items(items) }
+        assert_raise(ArgumentError) { set.set_data_items(nil) }
+        assert_raise(ArgumentError) { set.set_data_items([1]) }
+      end
+     
+      def test_get_mean_or_mode
+                items = [  [ "New York", 25, "Y"],
+                  [ "New York", 55, "Y"],
+                  [ "Chicago", 23, "Y"],
+                  [ "Boston", 23, "N"],
+                  [ "Chicago", 12, "N"],
+                  [ "Chicago", 87, "Y"] ]
+        set = DataSet.new.set_data_items(items)
+        assert_equal ["Chicago", 37.5, "Y"], set.get_mean_or_mode
+      end
+      
+      def test_index
+        items = [  [ "New York", 25, "Y"],
+                  [ "New York", 55, "Y"],
+                  [ "Chicago", 23, "Y"],
+                  [ "Boston", 23, "N"],
+                  [ "Chicago", 12, "N"],
+                  [ "Chicago", 87, "Y"] ]
+        set = DataSet.new.set_data_items(items)
+        assert_equal set.data_labels, set[0].data_labels
+        assert_equal [[ "New York", 25, "Y"]], set[0].data_items
+        assert_equal [[ "Chicago", 23, "Y"],[ "Boston", 23, "N"]], set[2..3].data_items
+        assert_equal items[1..-1], set[1..-1].data_items
+      end
+     
+    end
+  end
+end

data/test/data/proximity_test.rb

@@ -0,0 +1,81 @@
+# Author::    Sergio Fierens
+# License::   MPL 1.1
+# Project::   ai4r
+# Url::       http://ai4r.rubyforge.org/
+#
+# You can redistribute it and/or modify it under the terms of 
+# the Mozilla Public License version 1.1  as published by the 
+# Mozilla Foundation at http://www.mozilla.org/MPL/MPL-1.1.txt
+
+require 'test/unit'
+require File.dirname(__FILE__) + '/../../lib/ai4r/data/proximity'
+
+module Ai4r
+  module Data
+    class ProximityTest < Test::Unit::TestCase
+          
+      @@delta = 0.0001    
+      @@data1 = [rand*10, rand*10, rand*-10]
+      @@data2 = [rand*10, rand*-10, rand*10]
+      
+      def test_squared_euclidean_distance
+          assert_equal 0, Proximity.squared_euclidean_distance(@@data1, @@data1)
+          assert_equal  Proximity.squared_euclidean_distance(@@data1, @@data2), 
+                        Proximity.squared_euclidean_distance(@@data2, @@data1)
+          assert 0 <= Proximity.squared_euclidean_distance(@@data1, @@data1)
+          assert_equal 2, Proximity.squared_euclidean_distance([1,1], [2,2])
+          assert_equal 9, Proximity.squared_euclidean_distance([3], [0])
+      end
+      
+      def test_euclidean_distance
+          assert_equal 0, Proximity.euclidean_distance(@@data1, @@data1)
+          assert_equal  Proximity.euclidean_distance(@@data1, @@data2), 
+                        Proximity.euclidean_distance(@@data2, @@data1)
+          assert 0 <= Proximity.euclidean_distance(@@data1, @@data1)                        
+          assert_equal Math.sqrt(2), Proximity.euclidean_distance([1,1], [2,2])
+          assert_equal 3, Proximity.euclidean_distance([3], [0])
+      end
+
+      def test_manhattan_distance
+          assert_equal 0, Proximity.manhattan_distance(@@data1, @@data1)
+          assert_equal  Proximity.manhattan_distance(@@data1, @@data2), 
+                        Proximity.manhattan_distance(@@data2, @@data1)
+          assert 0 <= Proximity.manhattan_distance(@@data1, @@data1)                        
+          assert_equal 2, Proximity.manhattan_distance([1,1], [2,2])
+          assert_equal 9, Proximity.manhattan_distance([1,10], [2,2])
+          assert_equal 3, Proximity.manhattan_distance([3], [0])
+      end      
+      
+      def test_sup_distance
+          assert_equal 0, Proximity.sup_distance(@@data1, @@data1)
+          assert_equal  Proximity.sup_distance(@@data1, @@data2), 
+                        Proximity.sup_distance(@@data2, @@data1)
+          assert 0 <= Proximity.sup_distance(@@data1, @@data1)                        
+          assert_equal 1, Proximity.sup_distance([1,1], [2,2])
+          assert_equal 8, Proximity.sup_distance([1,10], [2,2])
+          assert_equal 3, Proximity.sup_distance([3], [0])
+      end  
+      
+      def test_hamming_distance
+          assert_equal 0, Proximity.hamming_distance(@@data1, @@data1)
+          assert_equal  Proximity.hamming_distance(@@data1, @@data2), 
+                        Proximity.hamming_distance(@@data2, @@data1)
+          assert 0 <= Proximity.hamming_distance(@@data1, @@data1)                        
+          assert_equal 1, Proximity.hamming_distance([1,1], [0,1])
+          assert_equal 2, Proximity.hamming_distance([1,10], [2,2])
+          assert_equal 1, Proximity.hamming_distance([3], [0])
+      end 
+      
+      def test_simple_matching_distance
+          assert_equal 0, Proximity.simple_matching_distance(@@data1, @@data1)
+          assert_equal  Proximity.simple_matching_distance(@@data1, @@data2), 
+                        Proximity.simple_matching_distance(@@data2, @@data1)
+          assert 0 <= Proximity.simple_matching_distance(@@data1, @@data1)                        
+          assert_equal 1, Proximity.simple_matching_distance([1,2], [0,1])
+          assert_equal 1.0/0, Proximity.simple_matching_distance([1,10], [2,2])
+          assert_equal 1.0/0, Proximity.simple_matching_distance([3], [0])
+      end       
+      
+    end
+  end
+end

data/test/data/statistics_data_set.csv

@@ -0,0 +1,5 @@
+1,2,3.0,4,5
+3,2,3.0,4,5
+6,2,3,4.0,5
+1.11,2,3.3,4,5
+1.4,2,3.7,4,5

data/test/data/statistics_test.rb

@@ -0,0 +1,65 @@
+# Author::    Sergio Fierens
+# License::   MPL 1.1
+# Project::   ai4r
+# Url::       http://ai4r.rubyforge.org/
+#
+# You can redistribute it and/or modify it under the terms of 
+# the Mozilla Public License version 1.1  as published by the 
+# Mozilla Foundation at http://www.mozilla.org/MPL/MPL-1.1.txt
+
+require 'test/unit'
+require File.dirname(__FILE__) + '/../../lib/ai4r/data/statistics'
+
+module Ai4r
+  module Data
+    class StatisticsTest < Test::Unit::TestCase
+      
+      DELTA = 0.00001
+      
+      def setup
+        @data_set = DataSet.new.
+          parse_csv "#{File.dirname(__FILE__)}/statistics_data_set.csv"
+      end
+    
+      def test_mean
+          assert_equal 2, Statistics.mean(@data_set, 1)
+          assert_equal 2.502, Statistics.mean(@data_set, 0)
+      end
+      
+      def test_variance
+          assert_equal 0, Statistics.variance(@data_set, 1)
+          assert_in_delta 4.47302, Statistics.variance(@data_set, 0), DELTA
+      end
+ 
+     def test_standard_deviation
+          assert_equal 0, Statistics.standard_deviation(@data_set, 1)
+          assert_in_delta 2.11495, Statistics.standard_deviation(@data_set, 0), DELTA
+      end
+ 
+      def test_mode
+        items = [ [ "New York", 25, "Y"],
+                  [ "New York", 55, "Y"],
+                  [ "Chicago", 23, "Y"],
+                  [ "Boston", 23, "N"],
+                  [ "Chicago", 12, "N"],
+                  [ "Chicago", 87, "Y"] ]
+        set = DataSet.new.set_data_items(items)
+        assert_equal "Chicago", Statistics.mode(set,0)
+        assert_equal 23, Statistics.mode(set,1)
+        assert_equal "Y", Statistics.mode(set,2)
+      end
+
+      def test_min
+          assert_equal 2, Statistics.min(@data_set, 1)
+          assert_equal 1, Statistics.min(@data_set, 0)
+      end
+      
+      def test_max
+          assert_equal 2, Statistics.max(@data_set, 1)
+          assert_equal 6, Statistics.max(@data_set, 0)
+          assert_equal 3.7, Statistics.max(@data_set, 2)
+      end      
+      
+    end
+  end
+end

data/test/experiment/classifier_evaluator_test.rb

@@ -0,0 +1,76 @@
+# Author::    Sergio Fierens
+# License::   MPL 1.1
+# Project::   ai4r
+# Url::       http://ai4r.rubyforge.org/
+#
+# You can redistribute it and/or modify it under the terms of 
+# the Mozilla Public License version 1.1  as published by the 
+# Mozilla Foundation at http://www.mozilla.org/MPL/MPL-1.1.txt
+
+require 'test/unit'
+require File.dirname(__FILE__) + '/../../lib/ai4r/experiment/classifier_evaluator'
+require File.dirname(__FILE__) + '/../../lib/ai4r/classifiers/classifier'
+
+class MockClassifier < Ai4r::Classifiers::Classifier
+  
+  attr_accessor :built
+  
+  def initialize(class_value)
+    @built = false
+    @class_value = class_value
+  end
+  
+  def build(data_set)
+    @built = true
+  end
+  
+  def eval(data)
+    @class_value
+  end
+  
+end
+
+module Ai4r
+  module Experiment
+    class ClassifierEvaluatorTest < Test::Unit::TestCase
+      
+      def test_add_classifier
+        evaluator = ClassifierEvaluator.new
+        evaluator << MockClassifier.new(nil) << MockClassifier.new(nil)
+        evaluator.add_classifier(MockClassifier.new(nil)).
+          add_classifier(MockClassifier.new(nil)).
+          add_classifier(MockClassifier.new(nil))
+        assert_equal 5, evaluator.classifiers.length
+      end
+
+      def test_build
+        evaluator = ClassifierEvaluator.new
+        5.times { evaluator << MockClassifier.new(nil) }
+        evaluator.classifiers.each {|c| assert !c.built}
+        evaluator.build(Ai4r::Data::DataSet.new)
+        evaluator.classifiers.each {|c| assert c.built}
+      end
+      
+      def test_eval
+        evaluator = ClassifierEvaluator.new
+        5.times { |x| evaluator << MockClassifier.new(x) }
+        assert_equal [0, 1, 2, 3, 4], evaluator.eval([])
+      end
+      
+      def test_test
+        evaluator = ClassifierEvaluator.new
+        5.times { |x| evaluator << MockClassifier.new(x) }
+        input = Ai4r::Data::DataSet.new :data_items => 
+          [[0],[0],[0],[1],[2],[3]]
+        output = evaluator.test input
+        assert_equal 5, output.data_items.length # 5 classifiers, 5 result rows
+        output.data_items.each { |result| assert result[1]>=0 && result[1]<1} # eval time
+        assert_equal 3, output.data_items.first[2] # 3 errors for the 1st classifier
+        assert_equal 0.5, output.data_items.first[3] # succes rate
+        assert_equal 6, output.data_items.last[2] # 6 errors for the last classifier
+        assert_equal 0, output.data_items.last[3] # succes rate
+      end
+      
+    end
+  end
+end