Module: ArrayMetrics

Defined in:
lib/array_metrics/pearson_correlation.rb

Class Method Summary collapse

Class Method Details

.constraint_pearson_correlation(x, y, m) ⇒ Object

Calculates the Constraint Pearson Correlation between Arrays x and y. See Shardanand and Maes 1995 for more details.

Shardanand, U. and Maes, P. Social information filtering: algorithms for automating “word of mouth”. In Proceedings of the SIGCHI conference on Human factors in computing systems (CHI ‘95). ACM Press/Addison-Wesley Publishing Co., New York, NY, USA, 210-217. DOI=10.1145/223904.223931 dx.doi.org/10.1145/223904.223931



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
 
# File 'lib/array_metrics/pearson_correlation.rb', line 10

def self.constraint_pearson_correlation(x, y, m)
  raise "array sizes don't match" if x.size != y.size

  cx, cy = [], []
  x.each_with_index do |xi, i|
    next if xi.nil? or y[i].nil?
    cx << xi
    cy << y[i]
  end

  return NaN if cx.size == 0

  s = sx = sy = 0.0
  if m.nil?
    cx_ = cx.mean
    cy_ = cy.mean
  else
    cx_ = cy_ = m
  end

  cx.each_with_index do |cxi,i|
    px = cxi-cx_
    py = cy[i]-cy_
    s  += px*py
    sx += px*px
    sy += py*py
  end

  s/Math.sqrt(sx*sy)
end

.pearson_correlation(x, y) ⇒ Object

Calculates the Pearson Correlation between the Arrays x and x. See secure.wikimedia.org/wikipedia/en/wiki/Pearson_correlation for more details.



44
45
46
 
# File 'lib/array_metrics/pearson_correlation.rb', line 44

def self.pearson_correlation(x,y)
  self.constraint_pearson_correlation(x,y,nil)
end