Class: RailsErrorDashboard::Services::PearsonCorrelation

Inherits:
Object
  • Object
show all
Defined in:
lib/rails_error_dashboard/services/pearson_correlation.rb

Overview

Pure algorithm: Calculate Pearson correlation coefficient between two series

No database access — accepts arrays, returns a correlation value. Used by ErrorCorrelation query to find time-correlated error types.

Examples:

PearsonCorrelation.call([1, 2, 3], [1, 2, 3])
# => 1.0 (perfect positive correlation)

Class Method Summary collapse

Class Method Details

.call(series_a, series_b) ⇒ Float

Returns Correlation coefficient between -1.0 and 1.0.

Parameters:

  • series_a (Array<Numeric>) —

    First data series

  • series_b (Array<Numeric>) —

    Second data series

Returns:

  • (Float) —

    Correlation coefficient between -1.0 and 1.0



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
43
 
# File 'lib/rails_error_dashboard/services/pearson_correlation.rb', line 17

def self.call(series_a, series_b)
  return 0.0 if series_a.empty? || series_b.empty?
  return 0.0 if series_a.sum.zero? || series_b.sum.zero?

  n = series_a.length
  return 0.0 if n.zero?

  mean_a = series_a.sum.to_f / n
  mean_b = series_b.sum.to_f / n

  covariance = 0.0
  std_a = 0.0
  std_b = 0.0

  n.times do |i|
    diff_a = series_a[i] - mean_a
    diff_b = series_b[i] - mean_b
    covariance += diff_a * diff_b
    std_a += diff_a**2
    std_b += diff_b**2
  end

  denominator = Math.sqrt(std_a * std_b)
  return 0.0 if denominator.zero?

  (covariance / denominator).round(3)
end