josephwilk / rsemantic

A document vector search with flexible matrix transforms. Currently supports Latent semantic analysis and Term frequency - inverse document frequency

This URL has Read+Write access

rsemantic / lib / semantic / transform / lsa_transform.rb
100644 42 lines (32 sloc) 1.296 kb
raw blame history 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

module Semantic
  module Transform
    class LSA
 
      class << self
 
        def transform(matrix, number_of_dimensions_to_reduce = 1)
          columns = matrix.num_columns
 
          if number_of_dimensions_to_reduce <= columns #Its a valid reduction
            u, sigma, vt = matrix.singular_value_decomposition
 
            sigma_prime = reduce_dimensions(number_of_dimensions_to_reduce, sigma)
 
            matrix_prime = u * sigma_prime * vt
          else
            raise Exception, "dimension reduction cannot be greater than %s" % columns
          end
          
          matrix_prime
        end
        
        private
        def reduce_dimensions(number_of_dimensions_to_reduce, matrix)
          for diagonal_index in dimensions_to_be_reduced(matrix, number_of_dimensions_to_reduce)
            matrix[diagonal_index, diagonal_index] = 0
          end
          matrix
        end
        
        def dimensions_to_be_reduced(matrix, number_of_dimensions_to_reduce)
          (diagonal_matrix_length(matrix) - number_of_dimensions_to_reduce)...diagonal_matrix_length(matrix)
        end
        
        def diagonal_matrix_length(matrix)
          matrix.num_columns < matrix.num_rows ? matrix.num_columns : matrix.num_rows
        end
        
      end
    end
  end
end