feature_selection 0.0.0 → 0.0.1

data/VERSION

@@ -1 +1 @@
-0.0.0
+0.0.1

data/benchmark/benchmark.rb

@@ -0,0 +1,17 @@
+require File.expand_path(File.dirname(__FILE__) + '/../lib/feature_selection')
+require 'benchmark'
+
+data = {:ham => [], :spam => []}
+
+1000.times do
+  a = rand(999).to_s
+  
+  data[:ham] << [a] * 5
+  data[:spam] << [a] * 5
+end
+
+Benchmark.bm do |x|
+  x.report do
+    FeatureSelection::MutualInformation.new(data).rank_features
+  end
+end

data/feature_selection.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{feature_selection}
-  s.version = "0.0.0"
+  s.version = "0.0.1"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["reddavis"]
-  s.date = %q{2010-01-05}
+  s.date = %q{2010-01-07}
   s.description = %q{A library of feature selection algorithms}
   s.email = %q{reddavis@gmail.com}
   s.extra_rdoc_files = [
@@ -23,6 +23,7 @@ Gem::Specification.new do |s|
      "README.rdoc",
      "Rakefile",
      "VERSION",
+     "benchmark/benchmark.rb",
      "feature_selection.gemspec",
      "lib/feature_selection.rb",
      "lib/feature_selection/base.rb",

data/lib/feature_selection/base.rb

@@ -13,54 +13,129 @@ module FeatureSelection
     
     # Contains term and belongs to class
     def n_1_1(term, klass)
-      count = 0.0
+      @n_1_1 = pre_compute_n_1_1 unless @n_1_1
+            
+      @n_1_1[klass][term]
+    end
+    
+    def pre_compute_n_1_1
+      results = {}
       
-      @data[klass].each do |document|
-        count += 1 if document.include?(term)
+      classes.each do |q_klass|
+        results[q_klass] = {}
+        
+        uniq_terms.each do |term|
+          count = 0.0
+
+          @data.each_pair do |klass, documents|
+            if klass == q_klass
+              documents.each do |document|
+                count += 1 if document.include?(term)
+              end
+            end
+          end
+
+          results[q_klass][term] = count
+        end
       end
-      
-      count
+
+      results
     end
     
     # Contains term but does not belong to class
     def n_1_0(term, klass)
-      count = 0.0
+      @n_1_0 = pre_compute_n_1_0 unless @n_1_0
+            
+      @n_1_0[klass][term]
+    end
+    
+    # Pre-Computer n_1_0
+    def pre_compute_n_1_0
+      results = {}
       
-      @data.each_pair do |key, documents|
-        if key != klass
-          documents.each do |document|
-            count += 1 if document.include?(term)
+      classes.each do |q_klass|
+        results[q_klass] = {}
+        
+        uniq_terms.each do |term|
+          count = 0.0
+
+          @data.each_pair do |klass, documents|
+            if klass != q_klass
+              documents.each do |document|
+                count += 1 if document.include?(term)
+              end
+            end
           end
+
+          results[q_klass][term] = count
         end
       end
-      
-      count
+
+      results
     end
     
     # Does not contain term but belongs to class
     def n_0_1(term, klass)
-      count = 0.0
+      @n_0_1 = pre_compute_n_0_1 unless @n_0_1
+            
+      @n_0_1[klass][term]
+    end
+    
+    # Pre-Computer n_0_1
+    def pre_compute_n_0_1
+      results = {}
       
-      @data[klass].each do |document|
-        count += 1 if !document.include?(term)
+      classes.each do |q_klass|
+        results[q_klass] = {}
+        
+        uniq_terms.each do |term|
+          count = 0.0
+
+          @data.each_pair do |klass, documents|
+            if klass == q_klass
+              documents.each do |document|
+                count += 1 if !document.include?(term)
+              end
+            end
+          end
+
+          results[q_klass][term] = count
+        end
       end
-      
-      count
+
+      results
     end
     
     # Does not contain term and does not belong to class
     def n_0_0(term, klass)
-      count = 0.0
+      @n_0_0 = pre_compute_n_0_0 unless @n_0_0
+            
+      @n_0_0[klass][term]
+    end
+    
+    # Pre-Computes all n_0_0 queries
+    def pre_compute_n_0_0
+      results = {}
       
-      @data.each_pair do |key, documents|
-        if key != klass
-          documents.each do |document|
-            count += 1 if !document.include?(term)
+      classes.each do |q_klass|
+        results[q_klass] = {}
+        
+        uniq_terms.each do |term|
+          count = 0.0
+
+          @data.each_pair do |klass, documents|
+            if klass != q_klass
+              documents.each do |document|
+                count += 1 if !document.include?(term)
+              end
+            end
           end
-        end #if key
-      end #@data.each_pair
-      
-      count
+
+          results[q_klass][term] = count
+        end
+      end
+
+      results
     end
     
     # All of the counts added together
@@ -76,9 +151,14 @@ module FeatureSelection
     def find_all_classes
       @data.map {|x| x[0]}
     end
+        
+    def uniq_terms
+      @uniq_terms ||= @data.map {|x| x[1]}.flatten.uniq
+    end
     
     def terms
-      @data.map {|x| x[1]}.flatten
+      @terms ||= @data.map {|x| x[1]}.flatten
     end
+    
   end
 end

data/lib/feature_selection/chi_squared.rb

@@ -13,7 +13,7 @@ module FeatureSelection
       classes.each do |klass|
         @results[klass] = {}
         
-        terms.each do |term|
+        uniq_terms.each do |term|
           answer = calculate_contribution(term, klass)
           @results[klass][term] = answer
         end #terms.each

data/lib/feature_selection/mutual_information.rb

@@ -24,7 +24,7 @@ module FeatureSelection
       classes.each do |klass|
         @results[klass] = {}
         
-        terms.each do |term|
+        uniq_terms.each do |term|
           answer = calculate_contribution(term, klass)
           @results[klass][term] = answer
         end #terms.each
@@ -49,28 +49,44 @@ module FeatureSelection
       begin
         if t == 1 && c == 1
           n_1_1 = n_1_1(term, klass)
-                    
+          
+          # return 0 if a == 0
+          a = (n * n_1_1) / ((n_0_1 + n_1_1) * (n_1_1 + n_1_0))
+          return 0.0 if a == 0
+          
           n_1_1 / n * 
-            Math.log( (n * n_1_1) / ((n_0_1 + n_1_1) * (n_1_1 + n_1_0)) ) 
+            Math.log(a) 
         elsif t == 1 && c == 0
           n_1_1 = n_1_1(term, klass)
           n_0_0 = n_0_0(term, klass)
-        
+          
+          # return 0 if a == 0
+          a = (n * n_1_0) / ((n_1_1 + n_0_1) * (n_0_1 + n_0_0))
+          return 0.0 if a == 0
+          
           n_1_0 / n *
-            Math.log( (n * n_1_0) / ((n_1_1 + n_0_1) * (n_0_1 + n_0_0)) )
+            Math.log(a)
         elsif t == 0 && c == 1
           n_0_0 = n_0_0(term, klass)
           n_1_1 = n_1_1(term, klass)
-        
+          
+          # return 0 if a == 0
+          a = (n * n_0_1) / ((n_1_0 + n_0_0) * (n_1_1 + n_1_0))
+          return 0.0 if a == 0
+          
           n_0_1 / n *
-            Math.log( (n * n_0_1) / ((n_1_0 + n_0_0) * (n_1_1 + n_1_0)) )
+            Math.log(a)
         elsif t == 0 && c == 0
           n_0_0 = n_0_0(term, klass)
+          
+          # return 0 if a == 0
+          a = (n * n_0_0) / ((n_1_0 + n_0_0) * (n_0_1 + n_0_0))
+          return 0.0 if a == 0
         
           n_0_0 / n *
-            Math.log( (n * n_0_0) / ((n_1_0 + n_0_0) * (n_0_1 + n_0_0)) )
+            Math.log(a)
         end
-      rescue ZeroDivisionError, Errno::EDOM
+      rescue ZeroDivisionError, Errno::EDOM #1.9 Infinity
         0.0
       end
     end

data/spec/feature_selection/chi_squared_spec.rb

@@ -9,5 +9,9 @@ describe "Chi Squared" do
   it "should return an hash" do
     @a.rank_features.should be_a(Hash)
   end
+  
+  it "should give this a score of 48.0" do
+    @a.rank_features[:spam]['this'].should == 48.0
+  end
     
 end

data/spec/feature_selection/mutual_information_spec.rb

@@ -9,5 +9,9 @@ describe "Mutual Information" do
   it "should return an hash" do
     @a.rank_features.should be_a(Hash)
   end
+  
+  it "should give this a score of 0.4904..." do
+    @a.rank_features[:spam]['this'].to_s.should match(/0.4904/)
+  end
     
 end

data/spec/spec_helper.rb

@@ -6,7 +6,7 @@ require 'spec/autorun'
 
 def data
   {
-  :spam => [['this', 'is', 'some', 'information'], ['this', 'is', 'something', 'that', 'is', 'information']],
+  :spam => [['this', 'is', 'some', 'yer', 'information'], ['this', 'is', 'something', 'that', 'is', 'information']],
   :ham => [['this', 'test', 'some', 'more', 'information'], ['there', 'are', 'some', 'things']],
   }
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: feature_selection
 version: !ruby/object:Gem::Version 
-  version: 0.0.0
+  version: 0.0.1
 platform: ruby
 authors: 
 - reddavis
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-01-05 00:00:00 +00:00
+date: 2010-01-07 00:00:00 +00:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -38,6 +38,7 @@ files:
 - README.rdoc
 - Rakefile
 - VERSION
+- benchmark/benchmark.rb
 - feature_selection.gemspec
 - lib/feature_selection.rb
 - lib/feature_selection/base.rb