multinomial_naive_bayes 0.1.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: e8ddc006ffc53986e48eae80b26a28bb4463328c
+  data.tar.gz: 096b99930d2f1fcbed9f599a3e7703b23791eea0
+SHA512:
+  metadata.gz: cb56810f0793160629bf2475e67b9aab4c5def427274a99bf1c65dc445a98dedd78008c02f3345ae2e9f1cd8a300f128185917c08fe894f661478c52f2997429
+  data.tar.gz: 1771a1e833359cf10aacc333ef1999daf8763311c7a784cbf8ca50db36bb1dfaccb648e4da625b94b9bb4c3c475054258bb9e70b8a87e2ef7235425c464b99a7

data/lib/multinomial_naive_bayes.rb

@@ -0,0 +1,2 @@
+require "multinomial_naive_bayes/classifier"
+require "multinomial_naive_bayes/learner"

data/lib/multinomial_naive_bayes/classifier.rb

@@ -0,0 +1,34 @@
+module MultinomialNaiveBayes
+  class Classifier
+    def initialize(categories_summaries, categories_probabilities)
+      @categories_summaries = categories_summaries
+      @categories_probabilities = categories_probabilities
+    end
+
+    def classify(vector)
+      max_ln_category_probability(vector)[0]
+    end
+
+    def max_ln_category_probability(vector)
+      all_ln_categories_probabilities(vector).
+        to_a.
+        sort_by{|ln_category_probability| -ln_category_probability[1]}.
+        first
+    end
+
+    def all_ln_categories_probabilities(vector)
+      @categories_summaries.keys.inject({}) do |map, category|
+        map[category] = ln_category_probability(vector, category)
+        map
+      end
+    end
+
+    def ln_category_probability(vector, category)
+      sum = 0
+      vector.each_with_index do |feature_value, feature|
+        sum += Math.log(@categories_summaries[category][feature]**feature_value)
+      end
+      sum + Math.log(@categories_probabilities[category])
+    end
+  end
+end

data/lib/multinomial_naive_bayes/learner.rb

@@ -0,0 +1,47 @@
+module MultinomialNaiveBayes
+  class Learner
+    def initialize(alpha = 1.0)
+      @alpha = alpha
+    end
+
+    def train(vector, category)
+      @category_to_feature_group ||= {}
+      @category_to_feature_group[category] ||= {}
+      vector.each_with_index do |feature_value, feature|
+        @category_to_feature_group[category][feature] ||= []
+        @category_to_feature_group[category][feature] << feature_value
+      end
+      @number_of_features ||= vector.length
+      @category_to_num_instances ||= Hash.new(0)
+      @category_to_num_instances[category] += 1
+    end
+
+    def classifier
+      Classifier.new(categories_summaries, categories_probabilities)
+    end
+
+    def categories_summaries
+      @category_to_feature_group.inject({}) do |map, (category, feature_group)|
+        map[category] = category_summary(feature_group)
+        map
+      end
+    end
+
+    def category_summary(feature_group)
+      total_count_of_all_features =  feature_group.values.flatten.reduce(&:+)
+      feature_group.inject({}) do |map, (feature, feature_values)|
+        total_count_of_feature = feature_values.reduce(&:+)
+        map[feature] = (total_count_of_feature + @alpha).to_f/(total_count_of_all_features + @number_of_features)
+        map
+      end
+    end
+
+    def categories_probabilities
+      total_instances = @category_to_num_instances.values.reduce(&:+)
+      @category_to_num_instances.inject({}) do |map, (category, num_instances)|
+        map[category] = num_instances.to_f/total_instances
+        map
+      end
+    end
+  end
+end

metadata

@@ -0,0 +1,46 @@
+--- !ruby/object:Gem::Specification
+name: multinomial_naive_bayes
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- An Le
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2016-04-24 00:00:00.000000000 Z
+dependencies: []
+description: 
+email: 
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/multinomial_naive_bayes.rb
+- lib/multinomial_naive_bayes/classifier.rb
+- lib/multinomial_naive_bayes/learner.rb
+homepage: https://github.com/lntan/multinomial_naive_bayes
+licenses:
+- MIT
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.6.3
+signing_key: 
+specification_version: 4
+summary: Implement the naive Bayes algorithm for multinomially distributed data
+test_files: []