Class: Daru::DataFrame

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

Instance Method Summary collapse

Instance Method Details

#create_indicator_vectors_for_categorical_vectors! ⇒ Object

Create for all non-numeric vectors 0-1-indicator columns. Returns the names/indices of the non-numeric vectors and the corresponding created new vectors.

Returns

A Hash where the keys are the names of all non-numeric vectors, and their values are Arrays containing the names of the corresponding 0-1 valued vectors

Arguments

  • for_model_without_intercept - if false (which is the default), then the indicator vector for the first level of every categorical variable will be excluded. Otherwise, all levels for the first categorical variable are considered, but the first level for all others are excuded.

Usage

df = Daru::DataFrame.new([(1..7).to_a, ['a','b','b','a','c','d','c']], order: ['int','char']) 
df.create_indicator_vectors_for_categorical_vectors!
# => {:char=>[:char_lvl_a, :char_lvl_b, :char_lvl_c, :char_lvl_d]}
df
# => #<Daru::DataFrame:70180517472080 @name = 75ddbda4-d4df-41b2-a41e-2f600764061b @size = 7>
           int       char char_lvl_a char_lvl_b char_lvl_c char_lvl_d 
  0          1          a        1.0        0.0        0.0        0.0 
  1          2          b        0.0        1.0        0.0        0.0 
  2          3          b        0.0        1.0        0.0        0.0 
  3          4          a        1.0        0.0        0.0        0.0 
  4          5          c        0.0        0.0        1.0        0.0 
  5          6          d        0.0        0.0        0.0        1.0 
  6          7          c        0.0        0.0        1.0        0.0



49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
 
# File 'lib/mixed_models/daru_methods.rb', line 49

def create_indicator_vectors_for_categorical_vectors!
  indices = Hash.new

  self.each_vector_with_index do |vec, name|
    # all non-numeric vectors are considered categorical data
    unless vec.type == :numeric
      levels = vec.to_a.uniq
      # sort the levels if possible
      begin
        levels.sort!
      rescue
        levels = vec.to_a.uniq
      end
      level_indices = Array.new
      levels.each do |l|
        ind = "#{name}_lvl_#{l}".to_sym
        col = Array.new
        vec.each { |e| e==l ? col.push(1.0) : col.push(0.0) }
        self[ind] = col
        level_indices.push(ind)
      end
      indices[name] = level_indices
    end
  end
   
  return indices
end

#to_nm(dtype: :float64, stype: :dense) ⇒ Object

Transform a Daru::DataFrame into a NMatrix

Arguments

  • dtype - the dtype of the returned NMatrix; defaults to float64

  • stype - the stype of the returned NMatrix; defaults to dense



12
13
14
15
16
17
 
# File 'lib/mixed_models/daru_methods.rb', line 12

def to_nm(dtype: :float64, stype: :dense)
  n, m = self.nrows, self.ncols
  data_array = Array.new 
  0.upto(n-1) { |i| data_array.concat(self.row[i].to_a) }
  return NMatrix.new([n,m], data_array, dtype: dtype, stype: stype)
end