From Wikipedia, the free encyclopedia
The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.
Examples of the Crystal Ball function.
The Crystal Ball function , named after the Crystal Ball Collaboration (hence the capitalized initial letters), is a probability density function (PDF) commonly used to model various lossy processes in high-energy physics such as Bremsstrahlung by electrons. It consists of a Gaussian core portion and a power-law low-end tail, below a certain threshold. The function itself and its first derivative are both continuous .
The Crystal Ball function is given by:
f
(
x
;
α
,
n
,
x
¯
,
σ
)
=
N
⋅
{
exp
(
−
(
x
−
x
¯
)
2
2
σ
2
)
,
for
x
−
x
¯
σ
>
−
α
A
⋅
(
B
−
x
−
x
¯
σ
)
−
n
,
for
x
−
x
¯
σ
⩽
−
α
,
{\displaystyle f(x;\alpha ,n,{\bar {x}},\sigma )=N\cdot {\begin{cases}\exp(-{\frac {(x-{\bar {x}})^{2}}{2\sigma ^{2}}}),&{\mbox{for }}{\frac {x-{\bar {x}}}{\sigma }}>-\alpha \\A\cdot (B-{\frac {x-{\bar {x}}}{\sigma }})^{-n},&{\mbox{for }}{\frac {x-{\bar {x}}}{\sigma }}\leqslant -\alpha \end{cases}},}
where
A
=
(
n
|
α
|
)
n
⋅
exp
(
−
|
α
|
2
2
)
{\displaystyle A=\left({\frac {n}{\left|\alpha \right|}}\right)^{n}\cdot \exp \left(-{\frac {\left|\alpha \right|^{2}}{2}}\right)}
,
B
=
n
|
α
|
−
|
α
|
{\displaystyle B={\frac {n}{\left|\alpha \right|}}-\left|\alpha \right|}
,
N
=
1
σ
(
C
+
D
)
{\displaystyle N={\frac {1}{\sigma (C+D)}}}
,
C
=
n
|
α
|
⋅
1
n
−
1
⋅
exp
(
−
|
α
|
2
2
)
{\displaystyle C={\frac {n}{\left|\alpha \right|}}\cdot {\frac {1}{n-1}}\cdot \exp \left(-{\frac {\left|\alpha \right|^{2}}{2}}\right)}
,
D
=
π
2
(
1
+
erf
(
|
α
|
2
)
)
{\displaystyle D={\sqrt {\frac {\pi }{2}}}\left(1+\operatorname {erf} \left({\frac {\left|\alpha \right|}{\sqrt {2}}}\right)\right)}
,
with the error function erf.
The parameters of the function (that are usually determined by a fit) are:
N
{\displaystyle N}
is a normalization factor (Skwarnicki 1986)
α
>
0
{\displaystyle \alpha >0}
defines the point where the PDF changes from a power-law to a Gaussian distribution
n
>
1
{\displaystyle n>1}
is the power of the power-law tail
x
¯
{\displaystyle {\bar {x}}}
and
σ
{\displaystyle \sigma }
are the mean and the standard deviation of the Gaussian
External links
J. E. Gaiser, Appendix-F Charmonium Spectroscopy from Radiative Decays of the J/Psi and Psi-Prime, Ph.D. Thesis , SLAC-R-255 (1982). (This is a 205-page document in .pdf form – the function is defined on p. 178.)
M. J. Oreglia, A Study of the Reactions psi prime --> gamma gamma psi, Ph.D. Thesis , SLAC-R-236 (1980), Appendix D.
T. Skwarnicki, A study of the radiative CASCADE transitions between the Upsilon-Prime and Upsilon resonances, Ph.D Thesis , DESY F31-86-02(1986), Appendix E.
Discrete univariate
with finite support with infinite support
Continuous univariate
supported on a bounded interval supported on a semi-infinite interval supported on the whole real line with support whose type varies
Mixed univariate
Multivariate (joint) Directional Degenerate and singular Families