Class: NBClass::Classifier

Inherits:
Object
  • Object
show all
Defined in:
lib/nb_class/classifier.rb

Constant Summary collapse

SMALL_PROBABILITY = 
0.000001

Instance Method Summary collapse

Constructor Details

#initialize ⇒ Classifier

Returns a new instance of Classifier.



10
11
12
13
14
15
16
17
18
19
 
# File 'lib/nb_class/classifier.rb', line 10

def initialize
  @category_examples = {}
  @occurrences = {}
  @global_occurrences = {}
  @probabilities = {}
  @global_probabilities = {}
  @category_occurrences = {}
  @total_elements = 0
  @total_phrases = 0;
end

Dynamic Method Handling

This class handles dynamic methods through the method_missing method

#method_missing(name, *args, &block) ⇒ Object 



103
104
105
106
107
108
109
 
# File 'lib/nb_class/classifier.rb', line 103

def method_missing(name, *args, &block)
  @category_examples[name] = PhraseArray.new
  define_singleton_method name do
    @category_examples[name]
  end
  public_send(name)
end

Instance Method Details

#categories ⇒ Object 



21
22
23
 
# File 'lib/nb_class/classifier.rb', line 21

def categories
  @category_examples.keys
end

#category_probability(category) ⇒ Object 



84
85
86
 
# File 'lib/nb_class/classifier.rb', line 84

def category_probability(category)
  @category_occurrences[category].to_f / @total_phrases.to_f
end

#category_probability_given_elements(params) ⇒ Object 



88
89
90
91
92
93
94
95
96
97
98
99
100
101
 
# File 'lib/nb_class/classifier.rb', line 88

def category_probability_given_elements(params)
  category = params[:category]
  elements = params[:elements]
  probability = 1.0
  elements.each do |element|
    element_probability_given_category = element_probability_given_category(category: category, element: element)
    element_global_probability = element_probability(element)
    element_probability = element_probability_given_category / element_global_probability
    probability *= element_probability
  end
  category_probability = category_probability(category)
  probability *= category_probability
  probability
end

#classify(elements) ⇒ Object 



39
40
41
42
43
44
45
46
47
48
49
50
51
 
# File 'lib/nb_class/classifier.rb', line 39

def classify(elements)
  elements = Utils.break_phrase_in_word_array(elements)
  max_category = nil
  max_probability = 0.0
  @occurrences.keys.each do |key|
    probability = category_probability_given_elements(category: key,elements: elements)
    if probability > max_probability
      max_probability = probability
      max_category = key
    end
  end
  max_category
end

#element_occurrence(params) ⇒ Object 



53
54
55
56
57
58
59
60
61
 
# File 'lib/nb_class/classifier.rb', line 53

def element_occurrence(params)
  category = params[:category]
  element = params[:element]
  if category
    @occurrences[category][element]
  else
    @global_occurrences[element]
  end
end

#element_probability(element) ⇒ Object 



75
76
77
78
79
80
81
82
 
# File 'lib/nb_class/classifier.rb', line 75

def element_probability(element)
  element_occurrence = @global_occurrences[element]
  unless element_occurrence.nil?
    element_occurrence.to_f / @total_elements
  else
    SMALL_PROBABILITY
  end
end

#element_probability_given_category(params) ⇒ Object 



63
64
65
66
67
68
69
70
71
72
73
 
# File 'lib/nb_class/classifier.rb', line 63

def element_probability_given_category(params)
  category = params[:category]
  element  = params[:element]
  element_occurrence_in_category = @occurrences[category][element]
  total_elements_in_category = @occurrences[category].map{ |item| item[1] }.inject{ |sum, item| sum + item }
  unless element_occurrence_in_category.nil?
    element_occurrence_in_category.to_f / total_elements_in_category.to_f
  else
    SMALL_PROBABILITY
  end
end

#train ⇒ Object 



25
26
27
28
29
30
31
32
33
34
35
36
37
 
# File 'lib/nb_class/classifier.rb', line 25

def train
  @category_examples.keys.each do |category|
    @category_examples[category].each do |phrase|
      phrase.each do |word|
        count_element_occurrence_in_category(category,word)
        count_element_occurrence_in_global(word)
        @total_elements += 1
      end
      count_category_occurrence(category)
      @total_phrases += 1
    end
  end
end