Class: Linkterm::Base

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

Instance Method Summary collapse

Constructor Details

#initialize(options = {}) ⇒ Base

Returns a new instance of Base.



72
73
74
75
76
 
# File 'lib/linkterm.rb', line 72

def initialize(options = {})
  @doc_dir = options[:doc_dir]

  @documents = Documents.new Dir.open(@doc_dir).reject {|f| /^\./ =~ f }.map {|f| "#{@doc_dir}/" + f }
end

Instance Method Details

#idf(word) ⇒ Object 



78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
 
# File 'lib/linkterm.rb', line 78

def idf(word)
  unless @documents_count
    all_count = Hash.new(0)
    counter = Counter.new(@documents.filenames.length, 'idf')
    @documents.each do |document|
      counter.display!
      words = Set.new []
      document.each {|paragraph|
        paragraph.each {|node|
          words << node.surface.downcase if node.category == '名詞' and
            node.surface.length > 1 and
            not /^\W+$/ =~ node.surface and
            not /^\d+$/ =~ node.surface
        }
      }
      words.each {|word| all_count[word] += 1 }
    end
    @documents_count = all_count
  end

  Math.log( @documents.filenames.length.to_f / @documents_count[word] )
end

#rule_table ⇒ Object 



101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
 
# File 'lib/linkterm.rb', line 101

def rule_table
  table = []
  counter = Counter.new(@documents.filenames.size, 'trans to table')
  #@documents.each(0..3) do |document|
  @documents.each() do |document|
    counter.display!
    document.each {|paragraph|
      hash = Hash.new 0
      paragraph.each {|node|
        hash[node.surface.downcase] += 1 if node.category == '名詞' and
          node.surface.length > 2 and
          not /^\W+$/ =~ node.surface and
          not /^\d+$/ =~ node.surface
      }
      word_count = hash
      # TF-IDF値の高いものみ取り出す
      word_count = word_count.sort_by {|word, count| count * idf(word) }.reverse[0...3].to_hash
      #word_count = word_count.select {|word, count| count * idf(word) > 1.5 }.to_hash

      word_count = word_count.map {|word, _| word }.sort
      table << word_count unless word_count.empty?
    }
  end

  # rule induction
  table = table.map {|line|
    arr = line.partitions {|term| (/^[a-z_]+$/ =~ term) ? :eng : (term.length > 10) ? :long : :short }
    arr << [] until arr.size >= 3
    arr
  }

  array = (0...table.first.length).to_a.parm
  counter = Counter.new(array.length, 'induce rule')
  rules = array.map {|c1, c2|
    counter.display!
    pattern = '[ [x,y] | x <- table.map{|u| u[c1] }.to_set, y <- table.map{|u| u[c2] }.to_set, !x.empty?, !y.empty? ]'.lisc_do binding
    pattern.map {|a1, a2|
      spt = table.select {|u| u[c1] == a1 and u[c2] == a2 }.length / table.select {|u| u[c1] == a1 }.length
      {:a1 => a1, :a2 => a2, :spt => spt}
    }
  }.flatten.select {|r| r[:spt] >= 1/2 }

  rules
end