machine_learner 0.0.2 → 0.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b1824a541ef0b505bf7bc812fc9a7661e9e5b64c
-  data.tar.gz: bcdfa290c18f7e108c458f997beebd976a29ea05
+  metadata.gz: bf87164a6191ea911d85afaa076ad0469fd906eb
+  data.tar.gz: 36ae6f7878759fd3454664dc198869bb48c52c92
 SHA512:
-  metadata.gz: 8c9752df583302b767fa3b06413d81ab3707fc99bdccc2e0106c19ccd7bdd17ca418a545d57c850afbce90a5c1be0b3f50a37a5f17250b7ff84130e063e3ef5f
-  data.tar.gz: e1e651f7eb1596c2beafdffeb07c5d508565160fa4eca3c5054eb2ec7235fef27f85b574ce4de4f96000c40a91ab8907a0e696453576c46ba13f0cb889cc4395
+  metadata.gz: e3febd94592309785e96681d892321ada9889731efe76c9bfad206679a9a1ef2756b2ecf8006849cd3f9279a18a04bb11d44652bcb90f8db5787ef16052b854b
+  data.tar.gz: 65acb7e1618cade8704b918a777c5d79d957cb03328dd70c62ee5007e61e9a0b94cf4884370fe6826be1f0b31750b18ac2b5d9e4c764efe430ff171eb8441ad8

data/Rakefile

@@ -1,7 +1,4 @@
 require "bundler/gem_tasks"
-require "rspec/core/rake_task"
-
-RSpec::Core::RakeTask.new(:spec)
 
 task :default => :spec
 

data/lib/machine_learner/adaboost.rb

@@ -95,17 +95,21 @@ module MachineLearner
       end
     end
     
+    # 識別結果の生の値を返す
+    # @param x 特徴空間
+    # @return [Float] 識別結果の生の値(-1..1)
+    def classify_raw(x)
+      [@learners, @alphas].transpose.reduce(0) { |score, (l, a)| score += l.classify(x) * a }
+    end
+    
     # 識別を行う
     # @param x 特徴空間
     # @return [Fixnum] 識別結果
     def classify(x)
-      score = [@learners, @alphas].transpose.reduce(0) { |score, (l, a)|
-        score += l.classify(x) * a 
-      }
-      score > 0 ? 1 : -1
+      classify_raw(x) > 0 ? 1 : -1
     end
     
-    # Learnerを表すログ
+    # Learnerを表す文字列
     def to_s
       [@learners, @alphas].transpose.map {|l, a| "(#{a.round(3)} * #{l})" }.join(" + ");
     end

data/lib/machine_learner/bayes.rb

@@ -8,13 +8,17 @@ module MachineLearner
   # ナイブベイズ学習器
   class BayesLearner < Learner
     
+    # コンストラクタ
+    def initialize
+    end
+    
     # データを元に学習を行う
     # @param datas [Array<DataSet>] トレーニングデータの配列
     # @return [Array<Boolean>] 識別結果の配列
     def learn(datas, ds = nil)
       @training = datas
       @ys = @training.map{|data| data.y }
-      @candidate_y = @ys.uniq.compact
+      @candidate_x = []
     end
     
     # 識別を行う
@@ -22,24 +26,24 @@ module MachineLearner
     # @return [Fixnum] 識別結果
     def classify(xs)
       # 最も尤度の高い候補 y を探す
-      max_y, max_likelihood = 0, -Float::INFINITY
-      @candidate_y.each do |y|
-        likelihood = p_y_x(y, xs)
-        max_y, max_likelihood = y, likelihood if max_likelihood < likelihood
-      end
-      return max_y
+      classify_raw(xs).max{ |x, y| x[1] <=> y[1] }[0]
+    end
+    
+    def classify_raw(xs)
+      candidate_y.map{|y| [y, p_y_x(y, xs)]}.to_h
     end
     
     private
   
     # P(Y | X) を計算
     def p_y_x(y, xs)
+      # P(Y | X) =  P(Y) * product(P(Xi | Y))
       p = Math.log(p_y(y))
       xs.each_with_index do |xi, i|
         next if xi.nil?
         p += Math.log(p_xi_y(xi, i, y))
       end
-      return p
+      return Math.exp(p)
     end
   
     # P(Xi | Y) を計算
@@ -56,14 +60,18 @@ module MachineLearner
     # P(Y) を計算
     def p_y(y)
       count = @ys.count(y)
-      return (count + 1).to_f / (@ys.size + @candidate_y.size + 1).to_f
+      return (count + 1).to_f / (@ys.size + candidate_y.size + 1).to_f
     end
   
     # xの候補を列挙
     def candidate_x(field)
-      @candidate_x ||= []
       @candidate_x[field] ||= @training.map{|data| data.x[field]}.uniq.compact
     end
+    
+    # yの候補を列挙
+    def candidate_y
+      @candidate_y ||= @ys.uniq.compact
+    end
   end
   
 end

data/lib/machine_learner/dataset.rb

@@ -5,12 +5,14 @@ module MachineLearner
   # 特徴空間xと識別結果yを持つ
   class DataSet
     attr_reader :x, :y
+    attr_accessor :id
     
     # コンストラクタ
     # @param x 特徴空間
     # @param y 識別結果
-    def initialize(x, y)
+    def initialize(x, y, id = nil)
       @x, @y = x, y
+      @id = id
     end
     
     # @return データセットを表現する文字列
@@ -22,3 +24,4 @@ module MachineLearner
 end
 
 # vim: set et ts=2 sts=2 sw=2:
+

data/lib/machine_learner/version.rb

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

data/machine_learner.gemspec

@@ -21,7 +21,6 @@ Gem::Specification.new do |spec|
 
   spec.add_development_dependency "bundler", "~> 1.6"
   spec.add_development_dependency "rake", "~> 10.0"
-  spec.add_development_dependency 'rspec', '~> 0'
 end
 
 # vim: set et ts=2 sts=2 sw=2:

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: machine_learner
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
 platform: ruby
 authors:
 - YutaTanaka
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-08-11 00:00:00.000000000 Z
+date: 2014-12-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -38,20 +38,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '10.0'
-- !ruby/object:Gem::Dependency
-  name: rspec
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0'
 description: |
   This is a library for machine learning.
   You can use AdaBoost and Naive Bayes easily.
@@ -104,4 +90,3 @@ summary: Library for machine learning.
 test_files:
 - spec/machine_learner_spec.rb
 - spec/spec_helper.rb
-has_rdoc: