Johnson (1949) Perkembangan sistem distribusi untuk menyocokkan distribusi yang tidak

diketahui dan mentransformasikannya menjadi normal, disebut transformasi Johnson. Johnson

transformasi terdiri dari tiga transformasi yaitu lognormal, unbounded, dan bounded, seperti pada
tabel 9
Tabel 9 sistem Johnson dengan transformasi yang sesuai
Johnson sistem Transformasi
SL (Lognormal) 𝑌(𝛾 ∗ , 𝛿, 𝜀) = 𝛾 ∗ + 𝛿 ln(𝑋 − 𝜀)
SU (Unbounded) 𝑋−𝜀
𝑌(𝛾, 𝛿, 𝜀, 𝜆) = 𝛾 ∗ + 𝛿 sinh−1 ( )
𝜆
SB (Bounded) 𝑋−𝜀
𝑌(𝛾, 𝛿, 𝜀, 𝜆) = 𝛾 ∗ + 𝛿 ln ( )
𝜆+𝜀−𝑋

Sistem Johnson memilih distribusinya sendiri dimana,

 Sistem lognormal meliputi dari keluarga lognormal
 Sistem Unbounded meliputi distribusi dari negative tak hingga ke tak hingga dari yang
bawah ke sisi atas, misalnya t dan distribusi normal
 Sistem Bounded meliputi distribusi

The transformation is dependent on four parameters where γ and δ indicates shape, λ

shape and ε location (George & Ramachandran, 2011). Slifker and Shapiro (1980)
presented a method to select a suitable Johnson system and estimate the four Johnson
parameters. The idea was to distinguish bounded from unbounded systems through
evaluating the tails of the unknown distribution. The selection algorithm contains the
following operations:
1. Chose a z-score (0<z<1) and create the four points ±z and ±3z. As a rule of
thumb, the chosen value of z should take greater values the larger the number of
observations.
2. Determine the probabilities Pζ, where ζ= {-3z, -z, z, 3z}, from the standard
normal table and let xζ be the corresponding percentiles of the data values.
3. Define the discriminant d, calculated as
(22)
where m = x3z – xz, n = x-z – xz and p = xz – x-z. The Johnson system is chosen
dependent on the value of the discriminant. If d is less than 0,999, the bounded system is
chosen. If d is greater than 1,001, the unbounded is chosen and for any value in between,
the lognormal system is chosen.

Untuk memilih sistem Johnson dibawah (1-3), di estiminasikan menjadi empat parameter sebagai
berikut:
1. Distribusi Johnson SL
 2𝑍
𝛿̂ = 𝑚
ln 𝑝

𝑚
𝑝 −1
𝛾̂ ∗ = 𝛿̂ ln 1
𝑚 2
𝑝(𝑝)
( )
𝑚
𝑥𝑧 + 𝑥−𝑧 𝑝 𝑝 1+
𝜀̂ = − 𝑚
2 2 −1
𝑝
2. Distribusi Johnson SU
2𝑍
𝛿̂ =
1 𝑚 𝑛
cosh−1 (2 ( 𝑝 + 𝑝))

𝑛 𝑚
𝑝− 𝑝
𝛾̂ ∗ = 𝛿̂ sinh−1 1
𝑚𝑛 2
2 ( 𝑝 𝑝 − 1)
( )
1
𝑚𝑛 2
2𝑝 ( 𝑝 𝑝 − 1)
𝜆̂ = 1
𝑚 𝑛 𝑚 𝑛 2
( 𝑝 + 𝑝 − 2) ( 𝑝 + 𝑝 + 2)
𝑛 𝑚
𝑥𝑧 + 𝑥−𝑧 𝑝 (𝑝 − 𝑝 )
𝜀̂ = + 𝑚 𝑛
2 2 ( 𝑝 + 𝑝 − 2)
3. Distribusi Johnson SB
𝑧
𝛿̂ = 1
1 𝑝 𝑝 2
cosh−1 (2 ((1 + 𝑚) (1 + 𝑛)) )
1
𝑝 𝑝 𝑝 𝑝 2
(𝑛 − 𝑚) ((1 + 𝑚) (1 + 𝑛) − 4)
𝛾̂ = 𝛿̂ sinh−1 𝑝𝑝
2 (𝑚 𝑛 − 1)
( )
1
2 2
𝑝 𝑝
𝑝 (((1 + ) (1 + ) − 2) − 4)
𝑚 𝑛
𝜆̂ = 𝑝𝑝
𝑚𝑛 −1
 𝑝 𝑝
𝑥𝑧 + 𝑥−𝑧 𝜆 𝑝 (𝑛 − 𝑚)
𝜀̂ = − +
2 2 2 ( 𝑝 𝑝 − 1)
𝑚𝑛
After estimating the Johnson parameters, the data set X can be transformed and for the transformed
data Y, the Anderson-Darling test is performed to evaluate normality. The procedure is repeated with a
feasible step size and number of iterations with a new z value. The Johnson transformation function
corresponding to the z value with highest p-value will be selected.