Module: Idhja22::Dataset::TreeMethods

Included in:
Idhja22::Dataset
Defined in:
lib/idhja22/dataset/tree_methods.rb

Instance Method Summary collapse

Instance Method Details

#entropy ⇒ Object 



16
17
18
19
20
 
# File 'lib/idhja22/dataset/tree_methods.rb', line 16

def entropy
  total = self.size
  return 1.0 if total < Idhja22.config.min_dataset_size
  category_counts.values.inject(0.0) { |ent, count|  prop = count.to_f/total.to_f; ent-prop*Math.log(prop,2)  }
end

#partition(attr_label) ⇒ Object 



4
5
6
7
8
9
10
11
12
13
14
 
# File 'lib/idhja22/dataset/tree_methods.rb', line 4

def partition(attr_label)
  groups = Hash.new([])
  data.each do |datum|
    groups[datum[attr_label]] += [datum]
  end
  output = Hash.new
  groups.each do |value, data|
    output[value] = Dataset.new(data, attribute_labels, category_label)
  end
  return output
end

#terminating? ⇒ Boolean

Returns:

  • (Boolean) 


22
23
24
 
# File 'lib/idhja22/dataset/tree_methods.rb', line 22

def terminating?
  (probability > Idhja22.config.termination_probability) || (probability < (1-Idhja22.config.termination_probability))
end