Module: Distribution::MathExtension::Erfc

Defined in:
lib/distribution/math_extension/erfc.rb

Overview

Error function from GSL-1.9, with epsilon information. Access it using Math.erfc_e

Constant Summary collapse

P =
[2.97886562639399288862,
7.409740605964741794425,
6.1602098531096305440906,
5.019049726784267463450058,
1.275366644729965952479585264,
0.5641895835477550741253201704]
Q =
[3.3690752069827527677,
9.608965327192787870698,
17.08144074746600431571095,
12.0489519278551290360340491,
9.396034016235054150430579648,
2.260528520767326969591866945,
1.0]

Class Method Summary collapse

Class Method Details

.erfc8(x) ⇒ Object



33
34
35
# File 'lib/distribution/math_extension/erfc.rb', line 33

def erfc8(x)
  erfc8_sum(x) * Math.exp(-x * x)
end

.erfc8_sum(x) ⇒ Object

Estimates erfc(x) valid for 8 < x < 100 erfc8_sum from GSL-1.9



25
26
27
28
29
30
31
# File 'lib/distribution/math_extension/erfc.rb', line 25

def erfc8_sum(x)
  num = P[5]
  4.downto(0) { |i| num = x * num + P[i] }
  den = Q[6]
  5.downto(0) { |i| den = x * den + Q[i] }
  num / den
end

.evaluate(x, with_error = false) ⇒ Object

gsl_sf_erfc_e



38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
# File 'lib/distribution/math_extension/erfc.rb', line 38

def evaluate(x, with_error = false)
  ax = x.abs
  e = nil

  if ax <= 1.0
    t = 2 * ax - 1
    e = ChebyshevSeries.evaluate(:erfc_xlt1, t, with_error)
  elsif ax <= 5.0
    ex2 = Math.exp(-x * x)
    t   = (ax - 3).quo(2)
    e   = ChebyshevSeries.evaluate(:erfc_x15, t, with_error)
    if with_error
      e[0] *= ex2
      e[1] = ex2 * (e[1] + 2 * x.abs * Float::EPSILON)
    else
      e *= ex2
    end
  elsif ax < 10.0
    exterm = Math.exp(-x * x) / ax
    t      = (2 * ax - 15).quo(5)
    e = ChebyshevSeries.evaluate(:erfc_x510, t, with_error)
    if with_error
      e[0] *= exterm
      e[1] = exterm * (e[1] + 2 * x.abs * Float::EPSILON + Float::EPSILON)
    else
      e *= exterm
    end
  else
    e8 = erfc8(ax)
    e = with_error ? [e8, (x * x + 1) * Float::EPSILON * e8.abs] : e8
  end

  result = x < 0 ? 2 - (with_error ? e.first : e) : (with_error ? e.first : e)
  with_error ? [result, e.last + 2 * Float::EPSILON * result.abs] : result
end