gaussian_naive_bayes 0.1.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 4ae196f08bdf0a291dcf4051bb3d06b6a3b5bee7
+  data.tar.gz: 9ba53c7eb6e2b94a15440994879fbbc7828a9a65
+SHA512:
+  metadata.gz: 71d87ccbbe53ceb24fcad8375192286410bac830c4f4f95d9efc684ad4777d007ece18ebe61dc5a33fb7c9202b6403860954b5229cb9dd30be456f6e109fd9d4
+  data.tar.gz: 64d41fbac231ea14e6097ed64c98ee20fd6a575b3a81ea51cf3335bbffe6fa3200f00e7c346a71f80e03f63514ac17bbc80869a619c9f57ef56f984edc503fca

data/lib/gaussian_naive_bayes.rb

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

data/lib/gaussian_naive_bayes/classifier.rb

@@ -0,0 +1,39 @@
+module GaussianNaiveBayes
+  class Classifier
+    attr_reader :categories_summaries, :categories_probabilities
+    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 += ln_normal_distribution(feature_value, @categories_summaries[category][feature][:mean], @categories_summaries[category][feature][:standard_deviation])
+      end
+      sum + Math.log(@categories_probabilities[category])
+    end
+
+    def ln_normal_distribution(x, mean, stdev)
+      Math.log(1.0/(stdev*Math.sqrt(2*Math::PI))) - ((x - mean)**2)/(2*(stdev**2))
+    end
+  end
+end

data/lib/gaussian_naive_bayes/learner.rb

@@ -0,0 +1,54 @@
+module GaussianNaiveBayes
+  class Learner
+    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
+      @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)
+      feature_group.inject({}) do |map, (feature, feature_values)|
+        map[feature] = {}
+        map[feature][:mean] = average(feature_values)
+        map[feature][:standard_deviation] = standard_deviation(feature_values)
+        map
+      end
+    end
+
+    def average(numbers)
+      numbers.reduce(&:+).to_f/numbers.length
+    end
+
+    def standard_deviation(numbers)
+      mean = average(numbers)
+      variance = numbers.inject(0) do |sum, number|
+        sum += (number - mean)**2
+      end.to_f/(numbers.length - 1)
+      Math.sqrt(variance)
+    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: gaussian_naive_bayes
+version: !ruby/object:Gem::Version
+  version: 0.1.1
+platform: ruby
+authors:
+- An Le
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2016-04-23 00:00:00.000000000 Z
+dependencies: []
+description: 
+email: 
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/gaussian_naive_bayes.rb
+- lib/gaussian_naive_bayes/classifier.rb
+- lib/gaussian_naive_bayes/learner.rb
+homepage: https://github.com/lntan/gaussian_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 Gaussian Naive Bayes algorithm for classification
+test_files: []