Weibull Distribution

Vijay Kumar

Department of Mathematics and Statistics

February, 2023
 Weibull Moments Median Mode

Weibull Distribution:

Let Y be a standard exponential random variable with probability density function (pdf)

f (y) = e−y , y > 0.

Define
X = σY 1/c + µ, σ > 0, c > 0.
The distribution of X is known as the Weibull distribution with shape parameter c > 0, scale
parameter σ > 0, and the location parameter µ. Its probability density is given by

c x − µ c−1 x−µ c
    
f (x; c, σ, µ) = exp − ; x > µ.
σ σ σ

The cumulative distribution function (cdf) is given by

x−µ c
  
F (x; c, σ, µ) = 1 − exp − ; x > µ.
σ
V. Kumar, DDU Gorakhpur University, Gorakhpur B.Sc.-I : Weibull distribution Feb. 2023 2/7
 Weibull Moments Median Mode

Weibull Distribution contd...

Let us denote the three-parameter distribution by Weibull (c, σ, µ).

F −1 (p|c, σ, µ) = µ + σ(−ln (1 − p))1/c

A two-parameter Weibull distribution is denoted by Weibull(c, σ).

c  x c−1 n  x c o
f (x; c, σ) = exp − ; x > 0.
σ σ σ
The cumulative distribution function (cdf) is given by
n  x c o
F (x; c, σ) = 1 − exp − ; x > 0.
σ
If σ = 1., we have the one parameter Weibull distribution.

V. Kumar, DDU Gorakhpur University, Gorakhpur B.Sc.-I : Weibull distribution Feb. 2023 3/7
 Weibull Moments Median Mode

Two-parameter Weibull distribution:

c  x c−1 n  x c o
PDF: f (x; c, σ) = exp − ; x > 0.
σ σ n   oσ
x c
CDF: F (x; c, σ) = 1 − exp − ; x > 0.
  σ
1
Mean = σ Γ 1 +
c
(     2 )
2 2
Variance = σ 2 Γ 1 + − Γ 1+
c c
 1/c
1
Mode = σ Γ 1 + ; c⩾1
c
Median = σ[ln (2)]1/c

V. Kumar, DDU Gorakhpur University, Gorakhpur B.Sc.-I : Weibull distribution Feb. 2023 4/7
 Weibull Moments Median Mode

Moments about the Origin:

Z∞
r
E (X ) = xr f (x) dx
0
Z∞
c  x c−1 n  x c o
= xr exp − dx
σ σ σ
0
Z∞  r Z∞
1/c r
= σy exp(−y) dy = σ y r/c exp(−y) dy
0 0
r
r

=σ Γ 1+
c
 r
µ′r = σ r Γ 1 +
c

V. Kumar, DDU Gorakhpur University, Gorakhpur B.Sc.-I : Weibull distribution Feb. 2023 5/7
 Weibull Moments Median Mode

Median:

ZM
M edian = f (x) dx
0
ZM  
c x c−1 n  x c o 1
exp − dx =
σ σ σ 2
0
c
(M
Z/σ)
1
exp(−y) dy =
2
0
1
1 − exp (−(M /σ)c ) =
2
M = σ(log 2)1/c

For three parameter we have M = µ + σ(log 2)1/c

V. Kumar, DDU Gorakhpur University, Gorakhpur B.Sc.-I : Weibull distribution Feb. 2023 6/7
 Weibull Moments Median Mode

Mode:
 n  x c o
′ d c  x c−1
f (x) = 0 = exp −
dx σ σ σ
c (c − 1) c−2 n  x  c o c c−1 c c−1 n  x c o
x exp − + x x exp − =0
σc σ σ c σ c σ
c c−1 n  x c o  (c − 1) cxc−1 
x exp − − =0
σc σ x σc
(c − 1) cxc−1
 
f (x) − =0
x σc
(c − 1) cxc−1
 
− =0
x σc
c − 1 1/c
 
xM ode = σ ; c>1
c

1/c
For three parameter we have xM ode = µ + σ c−1 c ; c>1

V. Kumar, DDU Gorakhpur University, Gorakhpur B.Sc.-I : Weibull distribution Feb. 2023 7/7