Normal distribution

This article is about the univariate normal distribution. 1 Definition

For normally distributed vectors, see Multivariate normal
distribution.
1.1 Standard normal distribution
“Bell curve” redirects here. For other uses, see Bell
curve (disambiguation).
The simplest case of a normal distribution is known as
the standard normal distribution. This is a special case
In probability theory, the normal (or Gaussian) distri- when μ=0 and σ=1, and it is described by this probability
bution is a very common continuous probability distribu- density function:
tion. Normal distributions are important in statistics and
are often used in the natural and social sciences to repre-
e− 2 x
1 2
sent real-valued random variables whose distributions are
not known.[1][2] ϕ(x) = √
2π
The normal distribution is useful because of the central √
The factor 1/ 2π in this expression ensures that the total
limit theorem. In its most general form, under some con-
area under the curve ϕ(x) is equal to one.[4] The ½ in the
ditions (which include finite variance), it states that aver-
exponent ensures that the distribution has unit variance
ages of random variables independently drawn from in-
(and therefore also unit standard deviation). This function
dependent distributions converge in distribution to the
is symmetric
√ around x=0, where it attains its maximum
normal, that is, become normally distributed when the
value 1/ 2π ; and has inflection points at +1 and −1.
number of random variables is sufficiently large. Phys-
ical quantities that are expected to be the sum of many Authors may differ also on which normal distribution
independent processes (such as measurement errors) of- should be called the “standard” one. Gauss defined the
ten have distributions that are nearly normal.[3] Moreover, standard normal as having variance σ 2 = 21 , that is
many results and methods (such as propagation of uncer-
tainty and least squares parameter fitting) can be derived
e−x
2
analytically in explicit form when the relevant variables
ϕ(x) = √
are normally distributed. π
The normal distribution is sometimes informally called Stigler[5] goes even further, defining the standard normal
the bell curve. However, many other distributions are with variance σ 2 = 1 :
(2π)
bell-shaped (such as the Cauchy, Student’s t, and logistic
distributions). The terms Gaussian function and Gaussian
bell curve are also ambiguous because they sometimes re-
ϕ(x) = e−πx
2

fer to multiples of the normal distribution that cannot be

directly interpreted in terms of probabilities.
The probability density of the normal distribution is: 1.2 General normal distribution
Every normal distribution is a version of the standard nor-
1 −
(x−µ)2 mal distribution whose domain has been stretched by a
f (x | µ, σ ) = √
2
e 2σ 2
2σ 2 π factor σ (the standard deviation) and then translated by μ
(the mean value):
Where:
( )
• µ is mean or expectation of the distribution (and also 1 x−µ
its median and mode). f (x | µ, σ) = ϕ .
σ σ
• σ is standard deviation The probability density must be scaled by 1/σ so that the
• σ is variance
2 integral is still 1.
If Z is a standard normal deviate, then X = Zσ + μ will
A random variable with a Gaussian distribution is said to have a normal distribution with expected value μ and stan-
be normally distributed and is called a normal deviate. dard deviation σ. Conversely, if X is a general normal

1
2 2 PROPERTIES

deviate, then Z = (X − μ)/σ will have a standard normal 2 Properties

distribution.
Every normal distribution is the exponential of a The normal distribution is the only absolutely continuous
quadratic function: distribution whose cumulants beyond the first two (i.e.,
other than the mean and variance) are zero. It is also the
continuous distribution with the maximum entropy for a
f (x) = eax2 +bx+c specified mean and variance.[9][10]
The normal distribution is a subclass of the elliptical dis-
where a is negative and c is b2 /(4a) + ln(−a/π)/2 . In tributions. The normal distribution is symmetric about its
this form, the mean value μ is −b/(2a), and the variance mean, and is non-zero over the entire real line. As such it
σ2 is −1/(2a). For the standard normal distribution, a is may not be a suitable model for variables that are inher-
−1/2, b is zero, and c is − ln(2π)/2 . ently positive or strongly skewed, such as the weight of
a person or the price of a share. Such variables may be
better described by other distributions, such as the log-
1.3 Notation normal distribution or the Pareto distribution.

The standard Gaussian distribution (with zero mean and The value of the normal distribution is practically zero
unit variance) is often denoted with the Greek letter ϕ when the value x lies more than a few standard deviations
(phi).[6] The alternative form of the Greek phi letter, φ, away from the mean. Therefore, it may not be an appro-
is also used quite often. priate model when one expects a significant fraction of
outliers—values that lie many standard deviations away
The normal distribution is also often denoted by N(μ, from the mean—and least squares and other statistical in-
σ2 ).[7] Thus when a random variable X is distributed nor- ference methods that are optimal for normally distributed
mally with mean μ and variance σ2 , we write variables often become highly unreliable when applied to
such data. In those cases, a more heavy-tailed distribution
should be assumed and the appropriate robust statistical
X ∼ N (µ, σ ). 2 inference methods applied.
The Gaussian distribution belongs to the family of
stable distributions which are the attractors of sums of
1.4 Alternative parameterizations independent, identically distributed distributions whether
or not the mean or variance is finite. Except for the Gaus-
Some authors advocate using the precision τ as the pa- sian which is a limiting case, all stable distributions have
rameter defining the width of the distribution, instead of heavy tails and infinite variance. It is one of the few dis-
the deviation σ or the variance σ2 . The precision is nor- tributions that are stable and that have probability density
mally defined as the reciprocal of the variance, 1/σ2 .[8] functions that can be expressed analytically, the others
The formula for the distribution then becomes being the Cauchy distribution and the Lévy distribution.

√
τ −τ (x−µ)2 2.1 Symmetries and derivatives
f (x) = e 2 .
2π
The normal distribution f(x), with any mean μ and any
This choice is claimed to have advantages in numerical
positive deviation σ, has the following properties:
computations when σ is very close to zero and simplify
formulas in some contexts, such as in the Bayesian infer-
ence of variables with multivariate normal distribution. • It is symmetric around the point x = μ, which is at
′ the same time the mode, the median and the mean
Also the reciprocal of the standard deviation τ = 1/σ of the distribution and it divides the data in half.[11]
might be defined as the precision and the expression of
the normal distribution becomes • It is unimodal: its first derivative is positive for x <
μ, negative for x > μ, and zero only at x = μ.

τ ′ −(τ ′ )2 (x−µ)2 • The area under the curve and over the x-axis is unity.
f (x) = √ e 2 .
2π
• Its density has two inflection points (where the sec-
According to Stigler, this formulation is advantageous ond derivative of f is zero and changes sign), located
because of a much simpler and easier-to-remember for- one standard deviation away from the mean, namely
mula, the fact that the pdf has unit height at zero, and at x = μ − σ and x = μ + σ.[11]
simple approximate formulas for the quantiles of the dis-
tribution. • Its density is log-concave.[11]
2.3 Fourier transform and characteristic function 3

• Its density is infinitely differentiable, indeed The central absolute moments coincide with plain mo-
supersmooth of order 2.[12] ments for all even orders, but are nonzero for odd orders.
For any non-negative integer p,
• Its second derivative f′′(x) is equal to its derivative
with respect to its variance σ2 .
{√ } p ( p+1 )
2
Furthermore, the density ϕ of the standard normal dis- E [|X| ] = σ (p−1)!!· ifpodd is 2 2Γ
p p π =σ ·p
√ 2
tribution (with μ = 0 and σ = 1) also has the following 1 ifpeven is π
properties:
The last formula is valid also for any non-integer p > −1.
• Its first derivative ϕ′(x) is −xϕ(x). When the mean μ is not zero, the plain and absolute mo-
ments can be expressed in terms of confluent hypergeo-
• Its second derivative ϕ′′(x) is (x2 − 1)ϕ(x) metric functions 1 F 1 and U.
• More generally, its n-th derivative ϕ(n) (x) is
(−1)n Hen(x)ϕ(x), where Hen is the nth (probabilist) ( )
√ p 1 1 1
Hermite polynomial. [13] E [X p
] = σ p
· (−i 2) U − p, , − (µ/σ) 2
,
2 2 2
• If we want to calculate the probability of a normally ( 1+p ) ( )
distributed variable X with known µ and σ . Then E [|X|p ] = σ p ·2 p2 Γ √ 2 1 F1 − 1 p, 1 , − 1 (µ/σ)2 .
we can calculate it by Standard Normal Distribution π 2 2 2
equivalent to Y = X−µ σ using probability table. These expressions remain valid even if p is not integer.
See also generalized Hermite polynomials.
2.1.1 Differential equation

• It satisfies the differential equation

The expectation of X conditioned on the event that X lies
in an interval [a,b] is given by
e−µ /(2σ
2 2
)
σ 2 f ′ (x) + f (x)(x − µ) = 0, f (0) = √
2σ 2 π f (b) − f (a)
E [X|a < X < b] = µ − σ
F (b) − F (a)
or
where f(x) and F(x) respectively are the density and the
cumulative distribution function of X.
√ −µ2 τ /2
′ τe
f (x)+τ f (x)(x−µ) = 0, f (0) = √ .
2π 2.3 Fourier transform and characteristic
function
2.2 Moments
The Fourier transform of a normal distribution f with
See also: List of integrals of Gaussian functions mean μ and deviation σ is[14]

The plain and absolute moments of a variable X are the ∫ ∞

f (x)e−itx dx = eiµt e− 2 (σt)
1 2
p p
expected values of X and |X| ,respectively. If the ex- ϕ̂(t) =
−∞
pected value μ of X is zero, these parameters are called
central moments. Usually we are interested only in mo- where i is the imaginary unit. If the mean μ is zero, the
ments with integer order p. first factor is 1, and the Fourier transform is also a normal
distribution on the frequency domain, with mean 0 and
If X has a normal distribution, these moments exist and
standard deviation 1/σ. In particular, the standard normal
are finite for any p whose real part is greater than −1. For
distribution ϕ (with μ=0 and σ=1) is an eigenfunction of
any non-negative integer p, the plain central moments are
the Fourier transform.
{ In probability theory, the Fourier transform of the prob-
p 0 ifpodd, is ability distribution of a real-valued random variable X is
E [X ] = called the characteristic function of that variable, and can
σ p (p − 1)!! ifpeven. is
be defined as the expected value of ei tX , as a function of
Here n!! denotes the double factorial, that is, the product the real variable t (the frequency parameter of the Fourier
of every number from n to 1 that has the same parity as transform). This definition can be analytically extended
n. to a complex-value parameter t.[15]
4 3 CUMULATIVE DISTRIBUTION FUNCTION

2.4 Moment and cumulant generating The complement of the standard normal CDF, Q(x) =
functions 1 − Φ(x) , is often called the Q-function, especially
in engineering texts.[16][17] It gives the probability that
The moment generating function of a real random vari- the value of a standard normal random variable X will
able X is the expected value of etX , as a function of the exceed x. Other definitions of the Q-function, all of
real parameter t. For a normal distribution with mean μ which are simple transformations of Φ , are also used
and deviation σ, the moment generating function exists occasionally.[18]
and is equal to The graph of the standard normal CDF Φ has 2-fold
rotational symmetry around the point (0,1/2); that is,
1 2 2
Φ(−x)∫ = 1 − Φ(x)∫ . Its antiderivative (indefinite in-
M (t) = ϕ̂(−it) = eµt e 2 σ t tegral) Φ(x) dx is Φ(x) dx = xΦ(x) + ϕ(x) .

The cumulant generating function is the logarithm of the

moment generating function, namely • The cumulative distribution function (CDF) of the
standard normal distribution can be expanded by
Integration by parts into a series:
1
g(t) = ln M (t) = µt + σ 2 t2
2
Since this is a quadratic polynomial in t, only the first two [ ]
1 x3 x5 x2n+1
cumulants are nonzero, namely the mean μ and the vari- Φ(x) = 0.5+ √ ·e−x /2 x +
2
+ + ··· + + ···
2π 3 3·5 (2n + 1)!!
ance σ2 .

where !! denotes the double factorial. As an example, the

following Pascal function approximates the CDF:
3 Cumulative distribution function
function CDF(x:extended):extended; var
The cumulative distribution function (CDF) of the stan- value,sum:extended; i:integer; begin sum:=x;
dard normal distribution, usually denoted with the capital value:=x; for i:=1 to 100 do begin
Greek letter Φ (phi), is the integral value:=(value*x*x/(2*i+1)); sum:=sum+value; end;
result:=0.5+(sum/sqrt(2*pi))*exp(-(x*x)/2); end;
∫ x
1
e−t
2
Φ(x) = √ /2
dt
2π −∞
3.1 Standard deviation and tolerance in-
In statistics one often uses the related error function, tervals
or erf(x), defined as the probability of a random vari-
able with normal distribution of mean 0 and variance 1/2 Main article: Tolerance interval
falling in the range [−x, x] ; that is About 68% of values drawn from a normal distribution

∫ x
1
e−t dt
2
erf(x) = √
π −x

These integrals cannot be expressed in terms of elemen-

tary functions, and are often said to be special functions.
However, many numerical approximations are known;
see below.
The two functions are closely related, namely

[ ( )]
1 x
Φ(x) = 1 + erf √
2 2
For a generic normal distribution f with mean μ and de- For the normal distribution, the values less than one standard de-
viation σ, the cumulative distribution function is viation away from the mean account for 68.27% of the set; while
two standard deviations from the mean account for 95.45%; and
three standard deviations account for 99.73%.
( ) [ ( )]
x−µ 1 x−µ
F (x) = Φ = 1 + erf √ are within one standard deviation σ away from the mean;
σ 2 σ 2
 5

about 95% of the values lie within two standard devia- However, one can define the normal distribution with zero
tions; and about 99.7% are within three standard devia- variance as a generalized function; specifically, as Dirac’s
tions. This fact is known as the 68-95-99.7 (empirical) “delta function” δ translated by the mean μ, that is f(x)
rule, or the 3-sigma rule. = δ(x−μ). Its CDF is then the Heaviside step function
More precisely, the probability that a normal deviate lies translated by the mean μ, namely
in the range μ − nσ and μ + nσ is given by
{
( ) F (x) = 0 ifx < µ
n 1 ifx ≥ µ
F (µ+nσ)−F (µ−nσ) = Φ(n)−Φ(−n) = erf √ ,
2
To 12 significant figures, the values for n = 1, 2, …, 6
are:[19] 5 Central limit theorem

3.2 Quantile function

Further information: Quantile function § Normal distri-
bution

The quantile function of a distribution is the inverse of the

cumulative distribution function. The quantile function
of the standard normal distribution is called the probit
function, and can be expressed in terms of the inverse
error function:

√ As the number of discrete events increases, the function begins to

Φ−1 (p) = 2 erf−1 (2p − 1), p ∈ (0, 1). resemble a normal distribution

For a normal random variable with mean μ and variance

σ2 , the quantile function is n=1 n=4
p(k) p(k)
0.18 1/6 0.18
0.16 0.16

√ 0.14
0.12
0.14
0.12 73 / 648
F −1 (p) = µ+σΦ−1 (p) = µ+σ 2 erf−1 (2p−1), p ∈ (0, 1).
0.10 0.10
0.08 0.08
0.05 0.05
The quantile Φ−1 (p) of the standard normal distribution 0.04
0.02
0.04
0.02
is commonly denoted as zp. These values are used in 0.00
123456 k
0.00
4 14 24 k

hypothesis testing, construction of confidence intervals n=2 n=5

p(k) p(k)
and Q-Q plots. A normal random variable X will exceed 0.18 1/6 0.18
0.16 0.16
μ + σzp with probability 1 − p; and will lie outside the in- 0.14 0.14

terval μ ± σzp with probability 2(1 − p). In particular, the 0.12

0.10
0.12
0.10
65 / 648

quantile z₀.₉₇₅ is 1.96; therefore a normal random variable 0.08

0.05
0.08
0.05
will lie outside the interval μ ± 1.96σ in only 5% of cases. 0.04 0.04
0.02 0.02
0.00 0.00
The following table gives the multiple n of σ such that X 2 7 12 k 5 17,18 30 k

will lie in the range μ ± nσ with a specified probability n=3

p(k)
p. These values are useful to determine tolerance interval 0.18
0.16
for sample averages and other statistical estimators with 0.14 1/8
0.12
normal (or asymptotically normal) distributions:[20] 0.10
0.08
0.05
0.04
0.02

4 Zero-variance limit 0.00

3 10,11 18 k

Comparison of probability density functions, p(k) for the sum of

In the limit when σ tends to zero, the probability density
f(x) eventually tends to zero at any x ≠ μ, but grows with-
out limit if x = μ, while its integral remains equal to 1.
Therefore, the normal distribution cannot be defined as
an ordinary function when σ = 0.
n fair 6-sided dice to show their convergence to a normal distri-
bution with increasing n, in accordance to the central limit theo-
rem. In the bottom-right graph, smoothed profiles of the previous
graphs are rescaled, superimposed and compared with a normal
distribution (black curve).
6 7 OPERATIONS ON NORMAL DEVIATES

Main article: Central limit theorem 6 Maximum entropy

The central limit theorem states that under certain (fairly Of all probability distributions over the reals with a spec-
common) conditions, the sum of many random variables ified mean μ and variance σ2 , the normal distribution
will have an approximately normal distribution. More N(μ, σ2 ) is the one with maximum entropy.[22] If X is a
specifically, where X1 , …, Xn are independent and iden- continuous random variable with probability density f(x),
tically distributed random variables with the same arbi- then the entropy of X is defined as[23][24][25]
2
trary distribution, zero
√ mean, and variance σ ; and Z is
their mean scaled by n ∫ ∞
H(X) = − f (x) log f (x)dx = 12 (1 + log(2σ 2 π))
−∞
( )
1∑
n
√ where f(x) log f(x) is understood to be zero whenever
Z= n Xi
n i=1 f(x) = 0. This functional can be maximized, subject to
the constraints that the distribution is properly normalized
Then, as n increases, the probability distribution of Z will and has a specified variance, by using variational calculus.
tend to the normal distribution with zero mean and vari- A function with two Lagrange multipliers is defined:
ance σ2 .
∫ ∞ ( ∫ ∞ ) ( ∫ ∞
The theorem can be extended to variables Xi that are not
independent and/or not identically distributed if certain L = f (x) ln(f (x)) dx−λ 0 1 − f (x) dx −λ σ 2
− f (x
−∞ −∞ −∞
constraints are placed on the degree of dependence and
the moments of the distributions. where f(x) is, for now, regarded as some function with
mean μ and standard deviation σ . At maximum entropy,
Many test statistics, scores, and estimators encountered
a small variation δf(x) about f(x) will produce a variation
in practice contain sums of certain random variables in
δL about L which is equal to zero:
them, and even more estimators can be represented as
sums of random variables through the use of influence
functions. The central limit theorem implies that those ∫ ∞
( )
statistical parameters will have asymptotically normal 0 = δL = δf (x) ln(f (x)) + 1 + λ0 + λ(x − µ)2 dx
−∞
distributions.
The central limit theorem also implies that certain distri- Since this must hold for any small δf(x), the term in
butions can be approximated by the normal distribution, brackets must be zero, and solving for f(x) yields:
for example:

f (x) = e−λ0 −1−λ(x−µ)

2

• The binomial distribution B(n, p) is approximately

normal with mean np and variance np(1−p) for large Using the constraint equations to solve for λ0 and λ yields
n and for p not too close to zero or one. the normal distribution:

• The Poisson distribution with parameter λ is approx-

imately normal with mean λ and variance λ, for large f (x, µ, σ) = √ 1 e− (x−µ)
2
2σ 2
values of λ.[21] 2σ 2 π

• The chi-squared distribution χ2 (k) is approximately

normal with mean k and variance 2k, for large k. 7 Operations on normal deviates
• The Student’s t-distribution t(ν) is approximately The family of normal distributions is closed under linear
normal with mean 0 and variance 1 when ν is large. transformations: if X is normally distributed with mean
μ and standard deviation σ, then the variable Y = aX + b,
Whether these approximations are sufficiently accurate for any real numbers a and b, is also normally distributed,
depends on the purpose for which they are needed, and with mean aμ + b and standard deviation |a|σ.
the rate of convergence to the normal distribution. It is Also if X1 and X2 are two independent normal random
typically the case that such approximations are less accu- variables, with means μ1 , μ2 and standard deviations
rate in the tails of the distribution. σ1 , σ2 , then their sum X1 + X2 will also be normally
[proof]
A general upper bound for the approximation error in the distributed, with mean μ1 + μ2 and variance σ12 +σ22
central limit theorem is given by the Berry–Esseen the- .
orem, improvements of the approximation are given by In particular, if X and Y are independent normal devi-
the Edgeworth expansions. ates with zero mean and variance σ2 , then X + Y and X
 7

− Y are also independent and normally distributed, with
zero mean and variance 2σ2 . This is a special case of the
polarization identity.[26]
Also, if X1 , X2 are two independent normal deviates with
mean μ and deviation σ, and a, b are arbitrary real num-
bers, then the variable
(named after Józef Marcinkiewicz) asserts that Q
can be at most a quadratic polynomial, and therefore
X a normal random variable.[29] The consequence of
this result is that the normal distribution is the only
distribution with a finite number (two) of non-zero
cumulants.

2. If X and Y are jointly normal and uncorrelated, then

they are independent. The requirement that X and
aX1 + bX2 − (a + b)µ Y should be jointly normal is essential, without it the
X3 = √ +µ
a2 + b2 property does not hold.[32][33][proof] For non-normal
random variables uncorrelatedness does not imply
is also normally distributed with mean μ and deviation
independence.
σ. It follows that the normal distribution is stable (with
exponent α = 2). 3. The Kullback–Leibler divergence of one normal
More generally, any linear combination of independent distribution X1 ∼ N(μ1 , σ2 1 )from another X2 ∼
normal deviates is a normal deviate. N(μ2 , σ2 2 )is given by:[34]
( )
(µ1 − µ2 )2 1 σ12 σ12
DKL (X1 ∥ X2 ) = + − 1 − ln .
7.1 Infinite divisibility and Cramér’s theo- 2σ22 2 σ22 σ22
rem The Hellinger distance between the same distribu-
tions is equal to
For any positive integer n, any normal distribution with
mean μ and variance σ2 is the distribution of the sum of
n independent normal deviates, each with mean μ/n and √ 2
variance σ2 /n. This property is called infinite divisibil- 2σ1 σ2 − 41 (µσ12−µ 2)
H (X1 , X2 ) = 1 −
2 2
e 1 +σ2 .
ity.[27] σ12 + σ22
Conversely, if X1 and X2 are independent random vari-
ables and their sum X1 + X2 has a normal distribution, 4. The Fisher information matrix for a normal distri-
then both X1 and X2 must be normal deviates.[28] bution is diagonal and takes the form
(1 )
This result is known as Cramér’s decomposition theo- 0
rem, and is equivalent to saying that the convolution of I= σ 2

0 2σ1 4
two distributions is normal if and only if both are nor-
mal. Cramér’s theorem implies that a linear combination 5. Normal distributions belongs to an exponential fam-
of independent non-Gaussian variables will never have an ily with natural parameters θ1 = σµ2 and θ2 = 2σ 2 , and
−1

exactly normal distribution, although it may approach it 2

natural statistics x and x . The dual, expectation pa-
arbitrarily closely.[29] rameters for normal distribution are η1 = μ and η2
= μ2 + σ 2 .
7.2 Bernstein’s theorem 6. The conjugate prior of the mean of a normal dis-
tribution is another normal distribution.[35] Specifi-
Bernstein’s theorem states that if X and Y are independent cally, if x1 , …, xn are iid N(μ, σ2 ) and the prior is μ
and X + Y and X − Y are also independent, then both X ~ N(μ0 , σ2
and Y must necessarily have normal distributions.[30][31] 0), then the posterior distribution for the estimator
More generally, if X1 , …, Xn are independent random of μ will be
variables, then two distinct linear combinations ∑akXk ( 2 ( )−1 )
σ 2
and ∑bkXk will be independent if and only if all Xk's are µ 0 + σ 0 x̄ n 1
µ | x1 , . . . , xn ∼ N n
σ2
, + 2
normal and ∑akbkσ 2 n + σ0
2 σ2 σ0
k = 0, where σ 2
k denotes the variance of Xk.[30] 7. The family of normal distributions forms a manifold
with constant curvature −1. The same family is
flat with respect to the (±1)-connections ∇(e) and
8 Other properties ∇(m) .[36]

1. If the characteristic function φX of some random

variable X is of the form φX(t) = eQ(t) , where Q(t) 9 Related distributions
is a polynomial, then the Marcinkiewicz theorem
8 9 RELATED DISTRIBUTIONS

9.1 Operations on a single random vari-

able
X12 + · · · + Xn2 ∼ χ2n .
2
If X is distributed normally with mean μ and variance σ ,
then • If X1 , X2 , …, Xn are independent normally dis-
tributed random variables with means μ and vari-
• The exponential of X is distributed log-normally: eX ances σ2 , then their sample mean is independent
~ ln(N (μ, σ2 )). from the sample standard deviation,[38] which can
be demonstrated using Basu’s theorem or Cochran’s
• The absolute value of X has folded normal distribu- theorem.[39] The ratio of these two quantities will
tion: |X| ~ Nf (μ, σ2 ). If μ = 0 this is known as the have the Student’s t-distribution with n − 1 degrees
half-normal distribution. of freedom:
• The absolute value of normalized residuals, |X -
μ|/σ, has chi distribution with one degree of free-
dom: |X - μ|/σ ~ χ1 (|X - μ|/σ).
X −µ n (X1 + · · · + Xn ) − µ
1
t= √ =√ [ ] ∼ tn−1 .
• The square of X/σ has the noncentral chi-squared S/ n 1
(X 1 − X) 2 + · · · + (X − X)2
n
distribution with one degree of freedom: X2 /σ2 ~ n(n−1)

χ2 1 (X2 /σ2 ). If μ = 0, the distribution is called sim-

ply chi-squared. • ' If X1 , …, Xn, Y 1 , …, Ym are independent standard
normal random variables, then the ratio of their nor-
• The distribution of the variable X restricted to an malized sums of squares will have the F-distribution
interval [a, b] is called the truncated normal distri- with (n, m) degrees of freedom:[40]
bution.
• (X − μ)−2 has a Lévy distribution with location 0 and
scale σ−2 . ( )
X12 + X22 + · · · + Xn2 /n
F = ∼ Fn, m .
(Y12 + Y22 + · · · + Ym2 ) /m
9.2 Combination of two independent ran-
dom variables
9.4 Operations on the density function
If X1 and X2 are two independent standard normal ran-
dom variables with mean 0 and variance 1, then The split normal distribution is most directly defined in
terms of joining scaled sections of the density functions
of different normal distributions and rescaling the density
• Their sum and difference is distributed normally
to integrate to one. The truncated normal distribution re-
with mean zero and variance two: X1 ± X2 ∼ N(0,
sults from rescaling a section of a single density function.
2).
• Their product Z = X1 ·X2 follows the “product-
normal” distribution[37] with density function fZ(z) 9.5 Extensions
= π−1 K 0 (|z|), where K 0 is the modified Bessel func-
tion of the second kind. This distribution is sym- The notion of normal distribution, being one of the most
metric around zero, unbounded at z = 0, and has the important distributions in probability theory, has been ex-
characteristic function φZ(t) = (1 + t 2 )−1/2 . tended far beyond the standard framework of the univari-
ate (that is one-dimensional) case (Case 1). All these ex-
• Their ratio follows the standard Cauchy distribution: tensions are also called normal or Gaussian laws, so a cer-
X1 ÷ X2 ∼ Cauchy(0, 1). tain ambiguity in names exists.
√ 2
• Their Euclidean norm X1 + X22 has the Rayleigh
distribution. • The multivariate normal distribution describes the
Gaussian law in the k-dimensional Euclidean space.
A vector X ∈ Rk is multivariate-normally distributed
9.3 Combination of two or more indepen- if any linear combination of its components ∑k
dent random variables j=1aj Xj has a (univariate) normal distribution. The
variance of X is a k×k symmetric positive-definite
• If X1 , X2 , …, Xn are independent standard normal matrix V. The multivariate normal distribution is a
random variables, then the sum of their squares has special case of the elliptical distributions. As such,
the chi-squared distribution with n degrees of free- its iso-density loci in the k = 2 case are ellipses and
dom in the case of arbitrary k are ellipsoids.
 9

• Rectified Gaussian distribution a rectified version of V (x) = (1 − 2 )(σ2 − σ1 )2 + σ1 σ2

normal distribution with all the negative elements π
√
reset to 0 2 4
T (x) = (σ2 − σ1 )[( − 1)(σ2 − σ1 )2 + σ1 σ2 ]
π π
• Complex normal distribution deals with the complex
k
normal vectors. A complex vector X ∈ C is said where E(x), V(x) and T(x) are the mean, variance and
to be normal if both its real and imaginary compo- third central moment respectively.
nents jointly possess a 2k-dimensional multivariate One of the main practical uses of the Gaussian law is to
normal distribution. The variance-covariance struc- model the empirical distributions of many different ran-
ture of X is described by two matrices: the variance dom variables encountered in practice. In such case a
matrix Γ, and the relation matrix C. possible extension would be a richer family of distribu-
tions, having more than two parameters and therefore be-
• Matrix normal distribution describes the case of ing able to fit the empirical distribution more accurately.
normally distributed matrices. The examples of such extensions are:
• Gaussian processes are the normally distributed
stochastic processes. These can be viewed as ele- • Pearson distribution— a four-parametric family of
ments of some infinite-dimensional Hilbert space H, probability distributions that extend the normal law
and thus are the analogues of multivariate normal to include different skewness and kurtosis values.
vectors for the case k = ∞. A random element h ∈
• The Generalized normal distribution, also known as
H is said to be normal if for any constant a ∈ H the
the exponential power distribution, allows for distri-
scalar product (a, h) has a (univariate) normal dis-
bution tails with thicker or thinner asymptotic be-
tribution. The variance structure of such Gaussian
haviors.
random element can be described in terms of the
linear covariance operator K: H → H. Several Gaus-
sian processes became popular enough to have their
own names: 10 Normality tests
• Brownian motion, Main article: Normality tests
• Brownian bridge,
• Ornstein–Uhlenbeck process. Normality tests assess the likelihood that the given data
set {x1 , …, xn} comes from a normal distribution. Typ-
• Gaussian q-distribution is an abstract mathematical ically the null hypothesis H 0 is that the observations are
construction that represents a "q-analogue" of the distributed normally with unspecified mean μ and vari-
normal distribution. ance σ2 , versus the alternative Ha that the distribution is
arbitrary. Many tests (over 40) have been devised for this
• the q-Gaussian is an analogue of the Gaussian dis- problem, the more prominent of them are outlined below:
tribution, in the sense that it maximises the Tsallis
entropy, and is one type of Tsallis distribution. Note
• “Visual” tests are more intuitively appealing but
that this distribution is different from the Gaussian
subjective at the same time, as they rely on infor-
q-distribution above.
mal human judgement to accept or reject the null
hypothesis.
A random variable x has a two piece normal distribution
if it has a distribution • Q-Q plot— is a plot of the sorted values from
the data set against the expected values of
the corresponding quantiles from the standard
normal distribution. That is, it’s a plot of
f (x) = N (µ, σ12 ) if x ≤ µ
point of the form (Φ−1 (pk), x₍k₎), where plot-
f (x) = N (µ, σ22 ) if x ≥ µ ting points pk are equal to pk = (k − α)/(n + 1
− 2α) and α is an adjustment constant, which
where μ is the mean and σ1 and σ2 are the standard devi- can be anything between 0 and 1. If the null
ations of the distribution to the left and right of the mean hypothesis is true, the plotted points should ap-
respectively. proximately lie on a straight line.
The mean, variance and third central moment of this dis- • P-P plot— similar to the Q-Q plot, but used
tribution have been determined[41] much less frequently. This method consists of
plotting the points (Φ(z₍k₎), pk), where z(k) =
(x(k) −µ̂)/σ̂ . For normally distributed data this
√
2 plot should lie on a 45° line between (0, 0) and
E(x) = µ + (σ2 − σ1 ) (1, 1).
π
10 11 ESTIMATION OF PARAMETERS

• Shapiro-Wilk test employs the fact that the

line in the Q-Q plot has the slope of σ. The
test compares the least squares estimate of that
slope with the value of the sample variance,
and rejects the null hypothesis if these two
quantities differ significantly.
• Normal probability plot (rankit plot)
• Moment tests:
• D'Agostino’s K-squared test
• Jarque–Bera test
• Empirical distribution function tests:
• Lilliefors test (an adaptation of
Kolmogorov–Smirnov test)
• Anderson–Darling test
µ̂ ∼ N (µ, σ 2/n).
The variance of this estimator is equal to the μμ-element
of the inverse Fisher information matrix I −1 . This im-
plies that the estimator is finite-sample efficient. Of prac-
tical importance is the fact that the standard error of µ̂ is
√
proportional to 1/ n , that is, if one wishes to decrease
the standard error by a factor of 10, one must increase
the number of points in the sample by a factor of 100.
This fact is widely used in determining sample sizes for
opinion polls and the number of trials in Monte Carlo
simulations.
From the standpoint of the asymptotic theory, µ̂ is
the consistent, that is, it converges in probability to μ as n →
∞. The estimator is also asymptotically normal, which
is a simple corollary of the fact that it is normal in finite
samples:

11 Estimation of parameters √ d
n(µ̂ − µ) −
→ N (0, σ 2 ).
See also: Maximum likelihood § Continuous distribu-
tion, continuous parameter space 11.2 Sample variance

It is often the case that we don't know the parameters See also: Standard deviation § Estimation, and Variance
of the normal distribution, but instead want to estimate § Estimation
them. That is, having a sample (x1 , …, xn) from a nor-
mal N(μ, σ2 ) population we would like to learn the ap- The estimator σ̂2 is called the sample variance, since it
proximate values of parameters μ and σ2 . The stan- is the variance of the sample (x1 , …, xn). In practice,
dard approach to this problem is the maximum likeli- another estimator is often used instead of the σ̂2 . This
hood method, which requires maximization of the log- other estimator is denoted s2 , and is also called the sam-
likelihood function: ple variance, which represents a certain ambiguity in ter-
minology; its square root s is called the sample standard
deviation. The estimator s2 differs from σ̂2 by having (n −
∑n
n n 1 ∑n
ln L(µ, σ 2 ) = ln f (xi ; µ, σ 2 ) = − ln(2π)− ln σ 2 −1) instead
2
of −µ)
(x i
n in the
2 denominator (the so-called Bessel’s
.
i=1
2 2 correction):
2σ i=1

Taking derivatives with respect to μ and σ2 and solving

1 ∑
n
the resulting system of first order conditions yields the 2 n
s = σ̂ 2 = (xi − x)2 .
maximum likelihood estimates: n−1 n − 1 i=1

The difference between s2 and σ̂2 becomes negligibly

1 ∑
n ∑
n
1 small for large n's. In finite samples however, the motiva-
µ̂ = x ≡ xi , σ̂ 2 = (xi − x)2 . tion behind the use of s2 is that it is an unbiased estima-
n i=1 n i=1
tor of the underlying parameter σ2 , whereas σ̂2 is biased.
Also, by the Lehmann–Scheffé theorem the estimator s2
11.1 Sample mean is uniformly minimum variance unbiased (UMVU),[42]
which makes it the “best” estimator among all unbiased
See also: Standard error of the mean ones. However it can be shown that the biased estimator
2
σ̂ 2 is “better” than the s in terms of the mean squared
error (MSE) criterion. In finite samples both s2 and σ̂2
Estimator µ̂ is called the sample mean, since it is the arith- have scaled chi-squared distribution with (n − 1) degrees
metic mean of all observations. The statistic x is complete of freedom:
and sufficient for μ, and therefore by the Lehmann–
Scheffé theorem, µ̂ is the uniformly minimum variance
unbiased (UMVU) estimator.[42] In finite samples it is σ2 σ2
distributed normally: s2 ∼ · χ2n−1 , σ̂ 2 ∼ · χ2n−1 .
n−1 n
 11

The first of these expressions shows that the variance of from the asymptotic distributions of µ̂ and s2 . The ap-
s2 is equal to 2σ4 /(n−1), which is slightly greater than the proximate formulas become valid for large values of n,
σσ-element of the inverse Fisher information matrix I −1 . and are more convenient for the manual calculation since
Thus, s2 is not an efficient estimator for σ2 , and moreover, the standard normal quantiles zα/2 do not depend on n.
since s2 is UMVU, we can conclude that the finite-sample In particular, the most popular value of α = 5%, results
efficient estimator for σ2 does not exist. in |z₀.₀₂₅| = 1.96.
Applying the asymptotic theory, both estimators s2 and
σ̂ 2 are consistent, that is they converge in probability to
σ2 as the sample size n → ∞. The two estimators are also 12 Bayesian analysis of the normal
both asymptotically normal: distribution
√ √ d Bayesian analysis of normally distributed data is compli-
n(σ̂ 2 − σ 2 ) ≃ n(s2 − σ 2 ) −
→ N (0, 2σ 4 ).
cated by the many different possibilities that may be con-
In particular, both estimators are asymptotically efficient sidered:
for σ2 .
• Either the mean, or the variance, or neither, may be
considered a fixed quantity.
11.3 Confidence intervals
• When the variance is unknown, analysis may be
See also: Studentization done directly in terms of the variance, or in terms
of the precision, the reciprocal of the variance. The
reason for expressing the formulas in terms of pre-
By Cochran’s theorem, for normal distributions the sam-
cision is that the analysis of most cases is simplified.
ple mean µ̂ and the sample variance s2 are independent,
which means there can be no gain in considering their • Both univariate and multivariate cases need to be
joint distribution. There is also a converse theorem: if in considered.
a sample the sample mean and sample variance are inde-
pendent, then the sample must have come from the nor- • Either conjugate or improper prior distributions may
mal distribution. The independence between µ̂ and s can be placed on the unknown variables.
be employed to construct the so-called t-statistic:
• An additional set of cases occurs in Bayesian lin-
ear regression, where in the basic model the data is
µ̂ − µ x−µ assumed to be normally distributed, and normal pri-
t= √ =√ ∑ ∼ tn−1 ors are placed on the regression coefficients. The
s/ n 1
(xi − x)2
n(n−1) resulting analysis is similar to the basic cases of
independent identically distributed data, but more
This quantity t has the Student’s t-distribution with (n − complex.
1) degrees of freedom, and it is an ancillary statistic (in-
dependent of the value of the parameters). Inverting the
distribution of this t-statistics will allow us to construct The formulas for the non-linear-regression cases are sum-
the confidence interval for μ;[43] similarly, inverting the marized in the conjugate prior article.
χ2 distribution of the statistic s2 will give us the confi-
dence interval for σ2 :[44]
12.1 Sum of two quadratics
[ ] [12.1.1 Scalar form ]
1 1 1 1
µ ∈ µ̂ − tn−1,1−α/2 √ s, µ̂ + tn−1,1−α/2 √ s ≈ µ̂ − |zα/2 | √ s, µ̂ + |zα/2 | √ s ,
n n n n
[ ] [ √ The following auxiliary
√ ] formula is useful for simplifying
(n − 1)s2
(n − 1)s2
2 the posterior update
2 equations, which otherwise become
σ2 ∈ , ≈ s2 − |zα/2 | √ s2 , s2 + |zα/2 | √ s2 ,
χ2n−1,1−α/2 χ2n−1,α/2 n fairly tedious. n

where tk,p and χ 2

( )2
k,p are the pth quantiles of the t- and χ2 -distributions ay + bz ab
respectively. These confidence intervals are of the a(x−y)2 +b(x−z)2 = (a+b) x − + (y−z)2
a+b a+b
confidence level 1 − α, meaning that the true values μ
and σ2 fall outside of these intervals with probability (or This equation rewrites the sum of two quadratics in x
significance level) α. In practice people usually take α by expanding the squares, grouping the terms in x, and
= 5%, resulting in the 95% confidence intervals. The completing the square. Note the following about the com-
approximate formulas in the display above were derived plex constant factors attached to some of the terms:
12 12 BAYESIAN ANALYSIS OF THE NORMAL DISTRIBUTION

a+b has the form of a weighted average N (µ, σ ) with known variance σ , the conjugate prior
1. The factor ay+bz 2 2

of y and z. distribution is also normally distributed.

2. ab
= 1+ 1
1 = (a
−1
+ b−1 )−1 . This shows that This can be shown more easily by rewriting the vari-
a+b a b
this factor can be thought of as resulting from a situ- ance as the precision, i.e. using τ = 1/σ2 . Then if
ation where the reciprocals of quantities a and b add x ∼ N (µ, 1/τ ) and µ ∼ N (µ0 , 1/τ0 ), we proceed as
directly, so to combine a and b themselves, it’s nec- follows.
essary to reciprocate, add, and reciprocate the re- First, the likelihood function is (using the formula above
sult again to get back into the original units. This for the sum of differences from the mean):
is exactly the sort of operation performed by the
ab
harmonic mean, so it is not surprising that a+b is
√ ( )
one-half the harmonic mean of a and b. ∏
n
τ 1
p(X | µ, τ ) = exp − τ (xi − µ)2
i=1
2π 2
( )
12.1.2 Vector form ( τ ) n2 1 ∑
n
= exp − τ (xi − µ) 2
2π 2 i=1
A similar formula can be written for the sum of two vector [ ( n )]
quadratics: If x, y, z are vectors of length k, and A and B ( τ ) n2 1 ∑
are symmetric, invertible matrices of size k × k , then = exp − τ (xi − x̄) + n(x̄ − µ)
2 2
.
2π 2 i=1

Then, we proceed as follows:

(y − x)′ A(y − x) + (x − z)′ B(x − z)
= (x − c)′ (A + B)(x − c) + (y − z)′ (A−1 + B−1 )−1 (y − z)
p(µ | X) ∝ p(X | µ)p(µ)
where [ ( n )] √
( τ ) n2 1 ∑ τ0
= exp − τ (xi − x̄) + n(x̄ − µ)
2 2
exp
2π 2 i=1
2π
c = (A + B)−1 (Ay + Bz) ( ( ( n ) )
1 ∑
∝ exp − τ (xi − x̄)2 + n(x̄ − µ)2 + τ0 (µ − µ0 )2
Note that the form x′ A x is called a quadratic form and 2 i=1
is a scalar: ( )
1( )
∝ exp − nτ (x̄ − µ)2 + τ0 (µ − µ0 )2
2
∑ ( ( )2
x′ Ax = aij xi xj 1 nτ x̄ + τ0 µ0 nτ τ0
= exp − (nτ + τ0 ) µ − + (x̄ − µ
i,j 2 nτ + τ0 nτ + τ0
( ( )2 )
In other words, it sums up all possible combinations of 1 nτ x̄ + τ0 µ0
products of pairs of elements from x, with a separate co- ∝ exp − (nτ + τ0 ) µ −
2 nτ + τ0
efficient for each. In addition, since xi xj = xj xi , only
the sum aij + aji matters for any off-diagonal elements In the above derivation, we used the formula above for the
of A, and there is no loss of generality in assuming that A sum of two quadratics and eliminated all constant factors
is symmetric. Furthermore, if A is symmetric, then the not involving μ. The result is the kernel of a normal dis-
form x′ Ay = y′ Ax. tribution, with mean nτ x̄+τ0 µ0 and precision nτ + τ0 ,
nτ +τ0
i.e.
12.2 Sum of differences from the mean
( )
nτ x̄ + τ0 µ0 1
Another useful formula is as follows: p(µ | X) ∼ N ,
nτ + τ0 nτ + τ0
This can be written as a set of Bayesian update equations
∑
n ∑
n
for the posterior parameters in terms of the prior param-
(xi − µ)2 = (xi − x̄)2 + n(x̄ − µ)2
i=1 i=1
eters:
1
∑n
where x̄ = n i=1 xi .
τ0′ = τ0 + nτ
nτ x̄ + τ0 µ0
12.3 With known variance µ′0 =
nτ + τ0
1∑
n
For a set of i.i.d. normally distributed data points X x̄ = xi
of size n where each individual point x follows x ∼ n i=1
12.5 With unknown mean and unknown variance 13

That is, to combine n data points with total precision of nτ

(or equivalently, total variance of n/σ2 ) and mean of val- ( ) n2 [ ]
1 ∑
n
1
ues x̄ , derive a new total precision simply by adding the p(X | µ, σ ) =2
exp − 2 (xi − µ)2

total precision of the data to the prior total precision, and 2πσ 2 2σ i=1
form a new mean through a precision-weighted average, ( ) n2 [ ]
1 S
i.e. a weighted average of the data mean and the prior = exp − 2
2πσ 2 2σ
mean, each weighted by the associated total precision.
This makes logical sense if the precision is thought of as where
indicating the certainty of the observations: In the distri-
bution of the posterior mean, each of the input compo-
∑n
nents is weighted by its certainty, and the certainty of this S = (xi − µ)2 .
distribution is the sum of the individual certainties. (For i=1
the intuition of this, compare the expression “the whole
Then:
is (or is not) greater than the sum of its parts”. In addi-
tion, consider that the knowledge of the posterior comes
from a combination of the knowledge of the prior and p(σ 2 | X) ∝ p(X | σ 2 )p(σ 2 )
likelihood, so it makes sense that we are more certain of [ ]
( ) n2 [ ] 2 ν0 ν0 exp −ν0 σ02
it than of either of its components.) 1 S (σ0 2 ) 2 2σ 2
= 2
exp − 2 ( ν0 ) ν0
The above formula reveals why it is more convenient 2πσ 2σ Γ 2 2
(σ ) 21+

to do Bayesian analysis of conjugate priors for the nor- ( ) n2 [ ]

1 1 S −ν0 σ02
mal distribution in terms of the precision. The posterior ∝ ν0 exp − +
σ2 (σ 2 )1+ 2 2σ 2 2σ 2
precision is simply the sum of the prior and likelihood [ ]
2
precisions, and the posterior mean is computed through 1 ν0 σ0 + S
= ν0 +n exp −
a precision-weighted average, as described above. The (σ 2 )1+ 2 2σ 2
same formulas can be written in terms of variance by re-
The above is also a scaled inverse chi-squared distribution
ciprocating all the precisions, yielding the more ugly for-
where
mulas

ν0′ = ν0 + n
′ 1
σ02 = ′ ∑
n
n
σ2 + 1
σ02 ν0′ σ02 = ν0 σ02 + (xi − µ)2
nx̄ µ0 i=1
σ2 + σ02
µ′0 = n 1
or equivalently
σ2 + σ02

1∑
n
x̄ = xi ν0′ = ν0 + n
n i=1 ∑n
′ ν0 σ02 + i=1 (xi − µ)2
σ02 =
ν0 + n
12.4 With known mean Reparameterizing in terms of an inverse gamma distribu-
tion, the result is:
For a set of i.i.d. normally distributed data points X
of size n where each individual point x follows x ∼ ′ n
N (µ, σ 2 ) with known mean μ, the conjugate prior of the α = α + 2
variance has an inverse gamma distribution or a scaled ∑n
i=1 (xi − µ)
2
′
inverse chi-squared distribution. The two are equivalent β = β + 2
except for having different parameterizations. Although
the inverse gamma is more commonly used, we use the
scaled inverse chi-squared for the sake of convenience. 12.5 With unknown mean and unknown
The prior for σ2 is as follows: variance
For a set of i.i.d. normally distributed data points X
[ ] [ ] of size n where each individual point x follows x ∼
−ν0 σ02 −ν0 σ02
(σ 2 ν0 ) 2 exp exp
ν0
2σ 2 N (µ, σ 2 ) with unknown mean μ and unknown variance
2σ 2
p(σ 2 | ν0 , σ02 ) = 0(2ν0 ) ν0 ∝ ν0 2
Γ 2 (σ 2 )1+ 2 (σ 2 )1+ 2 σ , a combined (multivariate) conjugate prior is placed
over the mean and variance, consisting of a normal-
The likelihood function from above, written in terms of inverse-gamma distribution. Logically, this originates as
the variance, is: follows:
14 12 BAYESIAN ANALYSIS OF THE NORMAL DISTRIBUTION

1. From the analysis of the case with unknown mean
but known variance, we see that the update equa-
tions involve sufficient statistics computed from the
data consisting of the mean of the data points and the
total variance of the data points, computed in turn
from the known variance divided by the number of
data points.
6. This leads immediately to the normal-inverse-
gamma distribution, which is the product of the two
distributions just defined, with conjugate priors used
(an inverse gamma distribution over the variance,
and a normal distribution over the mean, conditional
on the variance) and with the same four parameters
just defined.

2. From the analysis of the case with unknown variance

The priors are normally defined as follows:
but known mean, we see that the update equations
involve sufficient statistics over the data consisting
of the number of data points and sum of squared
deviations. p(µ | σ 2 ; µ0 , n0 ) ∼ N (µ0 , σ 2 /n0 )
p(σ 2 ; ν0 , σ02 ) ∼ Iχ2 (ν0 , σ02 ) = IG(ν0 /2, ν0 σ02 /2)
3. Keep in mind that the posterior update values serve
as the prior distribution when further data is han- The update equations can be derived, and look as follows:
dled. Thus, we should logically think of our priors in
terms of the sufficient statistics just described, with
the same semantics kept in mind as much as possi- 1∑
n

ble. x̄ = xi
n i=1
4. To handle the case where both mean and variance n0 µ0 + nx̄
µ′0 =
are unknown, we could place independent priors n0 + n
over the mean and variance, with fixed estimates of ′
n0 = n0 + n
the average mean, total variance, number of data ν0′ = ν0 + n
points used to compute the variance prior, and sum
′ ∑n
n0 n
of squared deviations. Note however that in reality, ν0′ σ02 = ν0 σ02 + (xi − x̄)2 + (µ0 − x̄)2
the total variance of the mean depends on the un- i=1
n 0 + n
known variance, and the sum of squared deviations
that goes into the variance prior (appears to) depend The respective numbers of pseudo-observations add the
on the unknown mean. In practice, the latter depen- number of actual observations to them. The new mean
dence is relatively unimportant: Shifting the actual hyperparameter is once again a weighted average, this
mean shifts the generated points by an equal amount, time weighted by the relative numbers of observations.
′
and on average the squared deviations will remain Finally, the update for ν0′ σ02 is similar to the case with
the same. This is not the case, however, with the known mean, but in this case the sum of squared devi-
total variance of the mean: As the unknown vari- ations is taken with respect to the observed data mean
ance increases, the total variance of the mean will rather than the true mean, and as a result a new “interac-
increase proportionately, and we would like to cap- tion term” needs to be added to take care of the additional
ture this dependence. error source stemming from the deviation between prior
and data mean.
5. This suggests that we create a conditional prior of the
[Proof]
mean on the unknown variance, with a hyperparam-
eter specifying the mean of the pseudo-observations
associated with the prior, and another parameter The prior distributions are
specifying the number of pseudo-observations. This
number serves as a scaling parameter on the vari-
1 ( n )
ance, making it possible to control the overall vari- 0
p(µ | σ 2 ; µ0 , n0 ) ∼ N (µ0 , σ 2 /n0 ) = √ exp − 2 (µ − µ0 )2
ance of the mean relative to the actual variance 2π nσ0
2 2σ
parameter. The prior for the variance also has ( n )
two hyperparameters, one specifying the sum of ∝ (σ 2 )−1/2 exp − 2 (µ − µ0 )2
0

squared deviations of the pseudo-observations asso- 2σ

ciated with the prior, and another specifying once p(σ 2
; ν0 , σ 2
0 ) ∼ Iχ 2
(ν 0 , σ 2
0 ) = IG(ν 2
0 /2, ν0 σ0 /2)
[ ]
again the number of pseudo-observations. Note that 2 ν0 /2 exp
−ν0 σ02
each of the priors has a hyperparameter specify- (σ ν0 /2) 2σ 2
= 0
ing the number of pseudo-observations, and in each Γ(ν0 /2) (σ ) 0 /2
2 1+ν
[ ]
case this controls the relative variance of that prior. 2 −(1+ν0 /2) −ν0 σ02
These are given as two separate hyperparameters so ∝ (σ ) exp
2σ 2
that the variance (aka the confidence) of the two pri-
ors can be controlled separately. Therefore, the joint prior is
13.1 Exact normality 15

2. Approximately normal laws, for example when such

approximation is justified by the central limit theo-
p(µ, σ 2 ; µ0 , n0 , ν0 , σ02 ) = p(µ | σ 2 ; µ0 , n0 ) p(σ 2 ; ν0 , σ02 ) rem; and
[ ]
2 −(ν0 +3)/2 1 ( )
∝ (σ ) exp − 2 ν0 σ03.+ Distributions
2
n0 (µ − µ0 ) modeled as normal – the normal distri-
2
2σ
bution being the distribution with maximum entropy
The likelihood function from the section above with for a given mean and variance.
known variance is:
4. Regression problems – the normal distribution be-
ing found after systematic effects have been mod-
( )n/2 [ ( n )] eled sufficiently well.
1 1 ∑
p(X | µ, σ 2 ) = 2
exp − 2 (xi − µ)2
2πσ 2σ i=1
13.1 Exact normality
Writing it in terms of variance rather than precision, we
get:

( )n/2 [ ( )]
1 1 ∑
n
p(X | µ, σ ) =
2
exp − 2 (xi − x̄)2 + n(x̄ − µ)2
2πσ 2 2σ i=1
[ ]
2 −n/2 1 ( )
∝σ exp − 2 S + n(x̄ − µ) 2
2σ
∑n
where S = i=1 (xi − x̄)2 .
Therefore, the posterior is (dropping the hyperparameters
as conditioning factors):

p(µ, σ 2 | X) ∝ p(µ, σ 2 ) p(X | µ, σ 2 )

[ ] [ ]
1 ( ) 2 −n/2 1 ( )
∝ (σ 2 )−(ν0 +3)/2 exp − 2 ν0 σ02 + n0 (µ − µ )2
σ exp −
The 0ground state of a quantum S + n(x̄ − µ) 2
2σ 2σ 2 harmonic oscillator has the
[ Gaussian distribution. ]
1 ( )
= (σ 2 )−(ν0 +n+3)/2 exp − 2 ν0 σ02 + S + n0 (µ − µ0 )2 + n(x̄ − µ)2
2σ Certain quantities in physics are distributed normally, as
[ ( ( Maxwell. Exam- )2 )]
1 was first
n0 ndemonstrated by James Clerk n0 µ0 + nx̄
2 −(ν0 +n+3)/2
= (σ ) exp − 2 ν0 σ0 + S ples
2
0 − x̄) +
+ of such(µquantities 2
(n0 + n) µ −
are:
2σ n0 + n n0 + n
[ ( )2 ]
n0 + n n0 µ0 + nx̄ • Probability density function of a ground state in a
∝ (σ 2 )−1/2 exp − µ−
2σ 2 n0 + n quantum harmonic oscillator.
[ ( )]
1 n0 n
× (σ 2 )−(ν0 /2+n/2+1) exp − 2 ν0 σ02 + S• +The position (µ0of−ax̄)
particle
2 that experiences diffusion.
2σ If ninitially
0 + n the particle is located at a specific point
( ) ( ( ))
n0 µ0 + nx̄ σ2 1 (that is
1 its probability distribution
n0 n is the2 dirac delta
= Nµ|σ2 , · IGσ2 (ν0 +function),
n), νthen
σ 2
+ S + (µ −
0 0 after time t its location
0 x̄) .
is described
n0 + n n0 + n 2 2 n0 + n
by a normal distribution with variance t, which sat-
In other words, the posterior distribution has the form of isfies the diffusion equation Template:Sfrac2 f(x,t)
a product of a normal distribution over p(μ | σ2 ) times an = Template:Sfrac2 Template:Sfrac2 f(x,t). If the
inverse gamma distribution over p(σ2 ), with parameters initial location is given by a certain density function
that are the same as the update equations above. g(x), then the density at time t is the convolution of
g and the normal PDF.

13 Occurrence and applications

13.2 Approximate normality
The occurrence of normal distribution in practical prob- Approximately normal distributions occur in many situ-
lems can be loosely classified into four categories: ations, as explained by the central limit theorem. When
the outcome is produced by many small effects acting ad-
1. Exactly normal distributions; ditively and independently, its distribution will be close to
16
normal. The normal approximation will not be valid if
the effects act multiplicatively (instead of additively), or
if there is a single external influence that has a consider-
ably larger magnitude than the rest of the effects.
• In counting problems, where the central limit the-
orem includes a discrete-to-continuum approxima-
tion and where infinitely divisible and decomposable
distributions are involved, such as
• Binomial random variables, associated with
binary response variables;
• Poisson random variables, associated with rare
events;
• Thermal radiation has a Bose–Einstein distribution
on very short time scales, and a normal distribution
on longer timescales due to the central limit theo-
rem.
13.3 Assumed normality
Histogram of sepal widths for Iris versicolor from Fisher’s Iris
flower data set, with superimposed best-fitting normal distribu-
tion.
I can only recognize the occurrence of the
normal curve – the Laplacian curve of errors –
as a very abnormal phenomenon. It is roughly
approximated to in certain distributions; for
this reason, and on account for its beautiful
simplicity, we may, perhaps, use it as a first
approximation, particularly in theoretical
investigations.
— Pearson (1901)
There are statistical methods to empirically test that as-
sumption, see the above Normality tests section.
• In biology, the logarithm of various variables tend
to have a normal distribution, that is, they tend to
have a log-normal distribution (after separation on
male/female subpopulations), with examples includ-
ing:
13 OCCURRENCE AND APPLICATIONS
• Measures of size of living tissue (length,
height, skin area, weight);[45]
• The length of inert appendages (hair, claws,
nails, teeth) of biological specimens, in the di-
rection of growth; presumably the thickness of
tree bark also falls under this category;
• Certain physiological measurements, such as
blood pressure of adult humans.
• In finance, in particular the Black–Scholes model,
changes in the logarithm of exchange rates, price in-
dices, and stock market indices are assumed normal
(these variables behave like compound interest, not
like simple interest, and so are multiplicative). Some
mathematicians such as Benoit Mandelbrot have ar-
gued that log-Levy distributions, which possesses
heavy tails would be a more appropriate model, in
particular for the analysis for stock market crashes.
The use of the assumption of normal distribution oc-
curring in financial models has also been criticized
by Nassim Nicholas Taleb in his works.
• Measurement errors in physical experiments are of-
ten modeled by a normal distribution. This use of a
normal distribution does not imply that one is as-
suming the measurement errors are normally dis-
tributed, rather using the normal distribution pro-
duces the most conservative predictions possible
given only knowledge about the mean and variance
of the errors.[46]
Fitted cumulative normal distribution to October rainfalls, see
distribution fitting
• In standardized testing, results can be made to have
a normal distribution by either selecting the number
and difficulty of questions (as in the IQ test) or trans-
forming the raw test scores into “output” scores by
fitting them to the normal distribution. For exam-
ple, the SAT's traditional range of 200–800 is based
on a normal distribution with a mean of 500 and a
standard deviation of 100.

• Many scores are derived from the normal distri-
bution, including percentile ranks (“percentiles” or
“quantiles”), normal curve equivalents, stanines,
z-scores, and T-scores. Additionally, some be-
havioral statistical procedures assume that scores
are normally distributed; for example, t-tests and
ANOVAs. Bell curve grading assigns relative grades
based on a normal distribution of scores.
• In hydrology the distribution of long duration river
discharge or rainfall, e.g. monthly and yearly totals,
is often thought to be practically normal according
to the central limit theorem.[47] The blue picture il-
lustrates an example of fitting the normal distribu-
tion to ranked October rainfalls showing the 90%
confidence belt based on the binomial distribution.
The rainfall data are represented by plotting posi-
tions as part of the cumulative frequency analysis.
13.4 Produced normality
In regression analysis, lack of normality in residuals sim-
ply indicates that the model postulated is inadequate in
accounting for the tendency in the data and needs to be
augmented; in other words, normality in residuals can al-
ways be achieved given a properly constructed model.
14 Generating values from normal
distribution
The bean machine, a device invented by Francis Galton, can be
called the first generator of normal random variables. This ma-
chine consists of a vertical board with interleaved rows of pins.
Small balls are dropped from the top and then bounce randomly
left or right as they hit the pins. The balls are collected into bins at
the bottom and settle down into a pattern resembling the Gaussian
curve.
In computer simulations, especially in applications of the
Monte-Carlo method, it is often desirable to generate val-
ues that are normally distributed. The algorithms listed
17
below all generate the standard normal deviates, since a
N(μ, σ2
) can be generated as X = μ + σZ, where Z is stan-
dard normal. All these algorithms rely on the availability
of a random number generator U capable of producing
uniform random variates.
• The most straightforward method is based on the
probability integral transform property: if U is dis-
tributed uniformly on (0,1), then Φ−1 (U) will have
the standard normal distribution. The drawback of
this method is that it relies on calculation of the
probit function Φ−1 , which cannot be done analyt-
ically. Some approximate methods are described in
Hart (1968) and in the erf article. Wichura gives
a fast algorithm for computing this function to 16
decimal places,[48] which is used by R to compute
random variates of the normal distribution.
• An easy to program approximate approach, that re-
lies on the central limit theorem, is as follows: gen-
erate 12 uniform U(0,1) deviates, add them all up,
and subtract 6 – the resulting random variable will
have approximately standard normal distribution. In
truth, the distribution will be Irwin–Hall, which is
a 12-section eleventh-order polynomial approxima-
tion to the normal distribution. This random deviate
will have a limited range of (−6, 6).[49]
• The Box–Muller method uses two independent ran-
dom numbers U and V distributed uniformly on
(0,1). Then the two random variables X and Y
√ √
X= −2 ln U cos(2πV ), Y = −2 ln U sin(2πV ).
will both have the standard normal distribu-
tion, and will be independent. This formula-
tion arises because for a bivariate normal ran-
dom vector (X, Y) the squared norm X2 + Y 2
will have the chi-squared distribution with two
degrees of freedom, which is an easily gener-
ated exponential random variable correspond-
ing to the quantity −2ln(U) in these equations;
and the angle is distributed uniformly around
the circle, chosen by the random variable V.
• Marsaglia polar method is a modification of the
Box–Muller method algorithm, which does not re-
quire computation of functions sin() and cos(). In
this method U and V are drawn from the uniform
(−1,1) distribution, and then S = U 2 + V 2 is com-
puted. If S is greater or equal to one then the method
starts over, otherwise two quantities
√ √
−2 ln S −2 ln S
X=U , Y =V
S S
18 15 NUMERICAL APPROXIMATIONS FOR THE NORMAL CDF

are returned. Again, X and Y will be indepen- The values Φ(x) may be approximated very accurately
dent and standard normally distributed. by a variety of methods, such as numerical integration,
Taylor series, asymptotic series and continued fractions.
• The Ratio method is a rejection method. The al- Different approximations are used depending on the de-
[50]

gorithm proceeds as follows: sired level of accuracy.

• Generate two independent uniform deviates U • A very simple and practical approximation is given
and V; by Bell [56] with a maximum absolute error of 0.003:
{ [ ]1/2 }
• Compute X = √8/e (V − 0.5)/U; 1 −2x2 /π
Φ(x) ≈ 1 + sign(x) 1 − e
• Optional: if X2 ≤ 5 − 4e1/4 U then accept X 2
and terminate algorithm; The inverse is also easily obtained. The inverse nor-
• Optional: if X ≥ 4e
2 −1.35
/U + 1.4 then reject mal formula is...
X and start over from step 1;
{
• If X2 ≤ −4 lnU then accept X, otherwise start (−1.57079632679 ln(1 − (2x − 1)2 ))1/2 , x > 0.5,
over the algorithm. Φ−1 (x) ≈
(−1.57079632679 ln(1 − (2x)2 ))1/2 , otherwise.

The two optional steps allow the evaluation of
the logarithm in the last step to be avoided
in most cases. These steps can be greatly
improved[51] so that the logarithm is rarely
evaluated.
• The ziggurat algorithm[52] is faster than the Box–
Muller transform and still exact. In about 97% of
all cases it uses only two random numbers, one ran-
dom integer and one random uniform, one multipli-
cation and an if-test. Only in 3% of the cases, where
the combination of those two falls outside the “core Φ(x) = 1−ϕ(x) b1 t + b2 t2 + b3 t3 + b4 t4 + b5 t5 +ε(x),
of the ziggurat” (a kind of rejection sampling using
logarithms), do exponentials and more uniform ran-
dom numbers have to be employed.
• Integer arithmetic can be used to sample from the
standard normal distribution.[53] This method is ex-
act in the sense that it satisfies the conditions of ideal
[54]
approximation; i.e., it is equivalent to sampling a
real number from the standard normal distribution
and rounding this to the nearest representable float-
ing point number.
• There is also some investigation[55] into the con-
nection between the fast Hadamard transform and
the normal distribution, since the transform employs
just addition and subtraction and by the central limit
theorem random numbers from almost any distribu-
tion will be transformed into the normal distribution.
In this regard a series of Hadamard transforms can
be combined with random permutations to turn ar-
bitrary data sets into a normally distributed data.
Φ(x) = +ϕ(x) x +
15 Numerical approximations for
the normal CDF
The standard normal CDF is widely used in scientific and
statistical computing.
Applying the following correction to the distribu-
tion’s tails...
if abs(.5-CDF)>.321 then do; if CDF>.5 then x=
1.0032*( x)^1.0362; else x=−1.0032*(-x)^1.0362;
end; guarantees an absolute relative error < 1.42% in
x, for CDF in [.00001,.99999]. All of this is easily
confirmed for a range of x of nearly 9 sigma.
• Zelen & Severo (1964) give the approximation for
Φ(x) for x > 0 with the absolute error | ε(x) | <
7.5·10−8 (algorithm 26.2.17):
( ) 1
t=
1+b
where ϕ(x) is the standard normal PDF, and b 0 =
0.2316419, b1 = 0.319381530, b2 = −0.356563782,
b3 = 1.781477937, b4 = −1.821255978, b5 =
1.330274429.
• Hart (1968) lists almost a hundred of rational func-
tion approximations for the erfc() function. His al-
gorithms vary in the degree of complexity and the
resulting precision, with maximum absolute preci-
sion of 24 digits. An algorithm by West (2009) com-
bines Hart’s algorithm 5666 with a continued frac-
tion approximation in the tail to provide a fast com-
putation algorithm with a 16-digit precision.
• Cody (1969) after recalling Hart68 solution is not
suited for erf, gives a solution for both erf and
erfc, with maximal relative error bound, via Rational
Chebyshev Approximation.
• Marsaglia (2004) suggested a simple algorithm[nb 1]
based on the Taylor series expansion
( )
1 x3 x5 x7 x9
+ + + + ···
2 3 3·5 3·5·7 3·5·7·9
for calculating Φ(x) with arbitrary precision. The
drawback of this algorithm is comparatively slow
calculation time (for example it takes over 300 it-
erations to calculate the function with 16 digits of
precision when x = 10).
16.1 Development 19

• The GNU Scientific Library calculates values of the tium" where among other things he introduces several im-
standard normal CDF using Hart’s algorithms and portant statistical concepts, such as the method of least
approximations with Chebyshev polynomials. squares, the method of maximum likelihood, and the nor-
mal distribution. Gauss used M, M′, M′′, … to denote the
Some more approximations can be found at: Error func- measurements of some unknown quantity V, and sought
tion#Approximation with elementary functions. the “most probable” estimator: the one that maximizes
the probability φ(M−V) · φ(M′−V) · φ(M′′−V) · … of
obtaining the observed experimental results. In his no-
tation φΔ is the probability law of the measurement er-
16 History rors of magnitude Δ. Not knowing what the function φ
is, Gauss requires that his method should reduce to the
16.1 Development well-known answer: the arithmetic mean of the measured
values.[nb 3] Starting from these principles, Gauss demon-
Some authors [57][58]
attribute the credit for the discovery strates that the only law that rationalizes the choice of
of the normal distribution to de Moivre, who in 1738[nb 2] arithmetic mean as an estimator of the location param-
[61]
published in the second edition of his "The Doctrine of eter, is the normal law of errors:
Chances" the study of the coefficients in the binomial ex- φ = √h e−hh∆∆ ,
π
pansion of (a + b)n . De Moivre proved that the middle
term where h is “the measure of the precision of the obser-
√ in this expansion has the approximate magnitude of
2/ 2πn , and that “If m or ½n be a Quantity infinitely vations”. Using this normal law as a generic model for
great, then the Logarithm of the Ratio, which a Term dis- errors in the experiments, Gauss formulates what is now
tant from the middle by the Interval ℓ, has to the middle known as the non-linear weighted least squares (NWLS)
[62]
Term, is − 2ℓℓn .”
[59]
Although this theorem can be inter- method.
preted as the first obscure expression for the normal prob-
ability law, Stigler points out that de Moivre himself did
not interpret his results as anything more than the approx-
imate rule for the binomial coefficients, and in particular
de Moivre lacked the concept of the probability density
function.[60]

Marquis de Laplace proved the central limit theorem in 1810,

consolidating the importance of the normal distribution in statis-
tics.

Although Gauss was the first to suggest the normal distri-

bution law, Laplace made significant contributions.[nb 4]
Carl Friedrich Gauss discovered the normal distribution in 1809
as a way to rationalize the method of least squares. It was Laplace who first posed the problem of aggregat-
ing several observations in 1774,[63] although his own so-
In 1809 Gauss published his monograph "Theoria motus lution led to the Laplacian distribution. It was Laplace
corporum coelestium in sectionibus conicis solem ambien- who first calculated the value of the integral ∫ e−t ² dt = √π
20 18 NOTES

in 1782, providing the normalization constant for the nor- cation parameter to the formula for normal distribution,
mal distribution.[64] Finally, it was Laplace who in 1810 expressing it in the way it is written nowadays:
proved and presented to the Academy the fundamental
central limit theorem, which emphasized the theoretical
importance of the normal distribution.[65] 1
e−(x−m) /(2σ ) dx
2 2
df = √
2
2σ π
It is of interest to note that in 1809 an American mathe-
matician Adrain published two derivations of the normal
The term “standard normal”, which denotes the normal
probability law, simultaneously and independently from
[66] distribution with zero mean and unit variance came into
Gauss. His works remained largely unnoticed by the
general use around the 1950s, appearing in the popular
scientific community, until in 1871 they were “rediscov-
[67] textbooks by P.G. Hoel (1947) "Introduction to mathe-
ered” by Abbe.
matical statistics" and A.M. Mood (1950) "Introduction
In the middle of the 19th century Maxwell demon- to the theory of statistics".[72]
strated that the normal distribution is not just a conve-
When the name is used, the “Gaussian distribution” was
nient mathematical tool, but may also occur in natural
[68] named after Carl Friedrich Gauss, who introduced the
phenomena: “The number of particles whose velocity,
distribution in 1809 as a way of rationalizing the method
resolved in a certain direction, lies between x and x + dx
of least squares as outlined above. Among English speak-
is
ers, both “normal distribution” and “Gaussian distribu-
tion” are in common use, with different terms preferred
by different communities.
1 x2
N √ e− α2 dx
α π

17 See also
16.2 Naming
• Behrens–Fisher problem — the long-standing prob-
Since its introduction, the normal distribution has been lem of testing whether two normal samples with dif-
known by many different names: the law of error, the ferent variances have same means;
law of facility of errors, Laplace’s second law, Gaussian
law, etc. Gauss himself apparently coined the term with • Bhattacharyya distance – method used to separate
reference to the “normal equations” involved in its ap- mixtures of normal distributions
plications, with normal having its technical meaning of
orthogonal rather than “usual”.[69] However, by the end • Erdős–Kac theorem—on the occurrence of the nor-
of the 19th century some authors[nb 5] had started using mal distribution in number theory
the name normal distribution, where the word “normal”
was used as an adjective – the term now being seen as a • Gaussian blur—convolution, which uses the normal
reflection of the fact that this distribution was seen as typ- distribution as a kernel
ical, common – and thus “normal”. Peirce (one of those
authors) once defined “normal” thus: "...the 'normal' is • Sum of normally distributed random variables
not the average (or any other kind of mean) of what ac-
tually occurs, but of what would, in the long run, occur • Normally distributed and uncorrelated does not im-
under certain circumstances.”[70] Around the turn of the ply independent
20th century Pearson popularized the term normal as a
• Tweedie distribution — The normal distribution is
designation for this distribution.[71]
a member of the family of Tweedie exponential dis-
persion models
Many years ago I called the Laplace–
Gaussian curve the normal curve, which • Z-test— using the normal distribution
name, while it avoids an international question
of priority, has the disadvantage of leading • Rayleigh distribution
people to believe that all other distributions
of frequency are in one sense or another • Multivariate normal distribution — a generalization
'abnormal'. of the normal distribution in multiple dimensions
— Pearson (1920)

18 Notes
Also, it was Pearson who first wrote the distribution in
terms of the standard deviation σ as in modern nota- [1] For example, this algorithm is given in the article Bc pro-
tion. Soon after this, in year 1915, Fisher added the lo- gramming language.

[2] De Moivre first published his findings in 1733, in a pam-
phlet “Approximatio ad Summam Terminorum Binomii
(a + b)n in Seriem Expansi” that was designated for pri-
vate circulation only. But it was not until the year 1738
that he made his results publicly available. The original
pamphlet was reprinted several times, see for example
Walker (1985).
[3] “It has been customary certainly to regard as an axiom
the hypothesis that if any quantity has been determined
by several direct observations, made under the same cir-
cumstances and with equal care, the arithmetical mean of
the observed values affords the most probable value, if not
rigorously, yet very nearly at least, so that it is always most
safe to adhere to it.” — Gauss (1809, section 177)
[4] “My custom of terming the curve the Gauss–Laplacian
or normal curve saves us from proportioning the merit of
discovery between the two great astronomer mathemati-
cians.” quote from Pearson (1905, p. 189)
[5] Besides those specifically referenced here, such use is en-
countered in the works of Peirce, Galton (Galton (1889,
chapter V)) and Lexis (Lexis (1878), Rohrbasser & Véron
(2003)) c. 1875.
19 Citations
[1] Normal Distribution, Gale Encyclopedia of Psychology
[2] Casella & Berger (2001, p. 102)
[3] Lyon, A. (2014). Why are Normal Distributions Nor-
mal?, The British Journal for the Philosophy of Science.
[4] For the proof see Gaussian integral
[5] Stigler (1982)
[6] Halperin, Hartley & Hoel (1965, item 7)
[7] McPherson (1990, p. 110)
[8] Bernardo & Smith (2000), p. 121
[9] Cover, Thomas M.; Thomas, Joy A. (2006). Elements of
Information Theory. John Wiley and Sons. p. 254.
[10] Park, Sung Y.; Bera, Anil K. (2009). “Maximum
Entropy Autoregressive Conditional Heteroskedasticity
Model” (PDF). Journal of Econometrics. Elsevier. 150
(2): 219–230. doi:10.1016/j.jeconom.2008.12.014. Re-
trieved 2011-06-02.
[11] Patel & Read (1996, [2.1.4])
[12] Fan (1991, p. 1258)
[13] Patel & Read (1996, [2.1.8])
[14] Bryc (1995, p. 23)
[15] Bryc (1995, p. 24)
[16] Scott, Clayton; Nowak, Robert (August 7, 2003). “The
Q-function”. Connexions.
21
[17] Barak, Ohad (April 6, 2006). “Q Function and Error
Function” (PDF). Tel Aviv University.
[18] Weisstein, Eric W. “Normal Distribution Function”.
MathWorld.
[19] WolframAlpha.com
[20] part 1, part 2
[21] Normal Approximation to Poisson(λ) Distribution, http:
//www.stat.ucla.edu/
[22] Cover & Thomas (2006, p. 254)
[23] Williams, D. (2001) Weighing the Odds Cambridge UP
ISBN 0-521-00618-X (pages 197-199)
[24] Bernardo, J.M., Smith, A.F.M. (2000) Bayesian Theory'.'
Wiley. ISBN 0-471-49464-X (pages 209, 366)
[25] O'Hagan, A. (1994) Kendall’s Advanced Theory of statis-
tics, Vol 2B, Bayesian Inference, Edward Arnold. ISBN
0-340-52922-9 (Section 5.40)
[26] Bryc (1995, p. 27)
[27] Patel & Read (1996, [2.3.6])
[28] Galambos & Simonelli (2004, Theorem 3.5)
[29] Bryc (1995, p. 35)
[30] Lukacs & King (1954)
[31] Quine, M.P. (1993). “On three characterisations of the
normal distribution”. Probability and Mathematical Statis-
tics. 14 (2): 257–263.
[32] UIUC, Lecture 21. The Multivariate Normal Distribution,
21.6:"Individually Gaussian Versus Jointly Gaussian”.
[33] Edward L. Melnick and Aaron Tenenbein, “Misspecifica-
tions of the Normal Distribution”, The American Statisti-
cian, volume 36, number 4 November 1982, pages 372–
373
[34] http://www.allisons.org/ll/MML/KL/Normal/
[35] Jordan, Michael I. (February 8, 2010). “Stat260:
Bayesian Modeling and Inference: The Conjugate Prior
for the Normal Distribution” (PDF).
[36] Amari & Nagaoka (2000)
[37] Normal Product Distribution, Mathworld
[38] Lukacs, Eugene (1942). “A Characterization of the Nor-
mal Distribution”. The Annals of Mathematical Statis-
tics. Institute of Mathematical Statistics. 13 (1): 91–
3. doi:10.1214/aoms/1177731647. ISSN 0003-4851.
JSTOR 2236166 – via JSTOR. (registration required
(help)).
[39] Basu, D.; Laha, R. G. (1954). “On Some Characteriza-
tions of the Normal Distribution”. Sankhyā. Indian Statis-
tical Institute. 13 (4): 359–62. ISSN 0036-4452. JSTOR
25048183 – via JSTOR. (registration required (help)).
[40] Lehmann, E. L. (1997). Testing Statistical Hypotheses
(2nd ed.). Springer. p. 199. ISBN 0-387-94919-4.
22
[41] John, S (1982). “The three parameter two-piece nor-
mal family of distributions and its fitting”. Communica-
tions in statistics - Theory and Methods. 11 (8): 879–885.
doi:10.1080/03610928208828279.
[42] Krishnamoorthy (2006, p. 127)
[43] Krishnamoorthy (2006, p. 130)
[44] Krishnamoorthy (2006, p. 133)
[45] Huxley (1932)
[46] Jaynes, Edwin T. (2003). Probability Theory: The Logic
of Science. Cambridge University Press. pp. 592–593.
[47] Oosterbaan, Roland J. (1994). “Chapter 6: Frequency and
Regression Analysis of Hydrologic Data”. In Ritzema,
Henk P. Drainage Principles and Applications, Publication
16 (PDF) (second revised ed.). Wageningen, The Nether-
lands: International Institute for Land Reclamation and
Improvement (ILRI). pp. 175–224. ISBN 90-70754-33-
9.
[48] Wichura, Michael J. (1988). “Algorithm AS241: The
Percentage Points of the Normal Distribution”. Ap-
plied Statistics. Blackwell Publishing. 37 (3): 477–84.
doi:10.2307/2347330. JSTOR 2347330 – via JSTOR.
(registration required (help)).
[49] Johnson, Kotz & Balakrishnan (1995, Equation (26.48))
[50] Kinderman & Monahan (1977)
[51] Leva (1992)
[52] Marsaglia & Tsang (2000)
[53] Karney (2016)
[54] Monahan (1985, section 2)
[55] Wallace (1996)
[56] Bell, Jeff (2015). “A Simple and Pragmatic Approxima-
tion to the Normal Cumulative Probability Distribution”.
SSRN Electronic Journal. doi:10.2139/ssrn.2579686.
[57] Johnson, Kotz & Balakrishnan (1994, p. 85)
[58] Le Cam & Lo Yang (2000, p. 74)
[59] De Moivre, Abraham (1733), Corollary I – see Walker
(1985, p. 77)
[60] Stigler (1986, p. 76)
[61] Gauss (1809, section 177)
[62] Gauss (1809, section 179)
[63] Laplace (1774, Problem III)
[64] Pearson (1905, p. 189)
[65] Stigler (1986, p. 144)
[66] Stigler (1978, p. 243)
[67] Stigler (1978, p. 244)
20 REFERENCES
[68] Maxwell (1860, p. 23)
[69] Jaynes, Edwin J.; Probability Theory: The Logic of Sci-
ence, Ch 7
[70] Peirce, Charles S. (c. 1909 MS), Collected Papers v. 6,
paragraph 327
[71] Kruskal & Stigler (1997)
[72] “Earliest uses… (entry STANDARD NORMAL
CURVE)".
20 References
• Aldrich, John; Miller, Jeff. “Earliest Uses of Sym-
bols in Probability and Statistics”.
• Aldrich, John; Miller, Jeff. “Earliest Known Uses of
Some of the Words of Mathematics”. In particular,
the entries for “bell-shaped and bell curve”, “normal
(distribution)", “Gaussian”, and “Error, law of error,
theory of errors, etc.”.
• Amari, Shun-ichi; Nagaoka, Hiroshi (2000). Meth-
ods of Information Geometry. Oxford University
Press. ISBN 0-8218-0531-2.
• Bernardo, José M.; Smith, Adrian F. M. (2000).
Bayesian Theory. Wiley. ISBN 0-471-49464-X.
• Bryc, Wlodzimierz (1995). The Normal Distribu-
tion: Characterizations with Applications. Springer-
Verlag. ISBN 0-387-97990-5.
• Casella, George; Berger, Roger L. (2001). Statistical
Inference (2nd ed.). Duxbury. ISBN 0-534-24312-
6.
• Cody, William J. (1969). "Rational Chebyshev
Approximations for the Error Function". Math-
ematics of Computation. 23 (107): 631–638.
doi:10.1090/S0025-5718-1969-0247736-4.
• Cover, Thomas M.; Thomas, Joy A. (2006). Ele-
ments of Information Theory. John Wiley and Sons.
• de Moivre, Abraham (1738). The Doctrine of
Chances. ISBN 0-8218-2103-2.
• Fan, Jianqing (1991). “On the optimal rates of
convergence for nonparametric deconvolution prob-
lems”. The Annals of Statistics. 19 (3): 1257–1272.
doi:10.1214/aos/1176348248. JSTOR 2241949.
• Galton, Francis (1889). Natural Inheritance (PDF).
London, UK: Richard Clay and Sons.
• Galambos, Janos; Simonelli, Italo (2004). Prod-
ucts of Random Variables: Applications to Problems
of Physics and to Arithmetical Functions. Marcel
Dekker, Inc. ISBN 0-8247-5402-6.

• Gauss, Carolo Friderico (1809). Theoria motvs cor-
porvm coelestivm in sectionibvs conicis Solem ambi-
entivm [Theory of the Motion of the Heavenly Bodies
Moving about the Sun in Conic Sections] (in Latin).
English translation.
• Gould, Stephen Jay (1981). The Mismeasure of Man
(first ed.). W. W. Norton. ISBN 0-393-01489-4.
• Halperin, Max; Hartley, Herman O.; Hoel, Paul
G. (1965). “Recommended Standards for Statis-
tical Symbols and Notation. COPSS Committee
on Symbols and Notation”. The American Statisti-
cian. 19 (3): 12–14. doi:10.2307/2681417. JSTOR
2681417.
• Hart, John F.; et al. (1968). Computer Approxi-
mations. New York, NY: John Wiley & Sons, Inc.
ISBN 0-88275-642-7.
• Hazewinkel, Michiel, ed. (2001), “Normal Dis-
tribution”, Encyclopedia of Mathematics, Springer,
ISBN 978-1-55608-010-4
• Herrnstein, Richard J.; Murray, Charles (1994).
The Bell Curve: Intelligence and Class Structure in
American Life. Free Press. ISBN 0-02-914673-9.
• Huxley, Julian S. (1932). Problems of Relative
Growth. London. ISBN 0-486-61114-0. OCLC
476909537.
• Johnson, Norman L.; Kotz, Samuel; Balakrish-
nan, Narayanaswamy (1994). Continuous Univari-
ate Distributions, Volume 1. Wiley. ISBN 0-471-
58495-9.
• Johnson, Norman L.; Kotz, Samuel; Balakrish-
nan, Narayanaswamy (1995). Continuous Univari-
ate Distributions, Volume 2. Wiley. ISBN 0-471-
58494-0.
• Karney, C. F. F. (2016). “Sampling exactly
from the normal distribution”. ACM Transac-
tions on Mathematical Software. 42 (1): 3:1–14.
arXiv:1303.6257 . doi:10.1145/2710016.
• Kinderman, Albert J.; Monahan, John F. (1977).
“Computer Generation of Random Variables Us-
ing the Ratio of Uniform Deviates”. ACM Trans-
actions on Mathematical Software. 3 (3): 257–260.
doi:10.1145/355744.355750.
• Krishnamoorthy, Kalimuthu (2006). Handbook of
Statistical Distributions with Applications. Chapman
& Hall/CRC. ISBN 1-58488-635-8.
• Kruskal, William H.; Stigler, Stephen M. (1997).
Spencer, Bruce D., ed. Normative Terminology:
'Normal' in Statistics and Elsewhere. Statistics and
Public Policy. Oxford University Press. ISBN 0-
19-852341-6.
23
• Laplace, Pierre-Simon de (1774). “Mémoire sur
la probabilité des causes par les événements”. Mé-
moires de l'Académie royale des Sciences de Paris
(Savants étrangers), tome 6: 621–656. Translated by
Stephen M. Stigler in Statistical Science 1 (3), 1986:
JSTOR 2245476.
• Laplace, Pierre-Simon (1812). Théorie analytique
des probabilités [Analytical theory of probabilities].
• Le Cam, Lucien; Lo Yang, Grace (2000). Asymp-
totics in Statistics: Some Basic Concepts (second ed.).
Springer. ISBN 0-387-95036-2.
• Leva, Joseph L. (1992). “A fast normal ran-
dom number generator”. ACM Transactions
on Mathematical Software. 18 (4): 449–453.
doi:10.1145/138351.138364.
• Lexis, Wilhelm (1878). “Sur la durée normale de
la vie humaine et sur la théorie de la stabilité des
rapports statistiques”. Annales de démographie in-
ternationale. Paris. II: 447–462.
• Lukacs, Eugene; King, Edgar P. (1954). “A
Property of Normal Distribution”. The Annals
of Mathematical Statistics. 25 (2): 389–394.
doi:10.1214/aoms/1177728796. JSTOR 2236741.
• McPherson, Glen (1990). Statistics in Scientific In-
vestigation: Its Basis, Application and Interpretation.
Springer-Verlag. ISBN 0-387-97137-8.
• Marsaglia, George; Tsang, Wai Wan (2000). “The
Ziggurat Method for Generating Random Vari-
ables”. Journal of Statistical Software. 5 (8).
doi:10.18637/jss.v005.i08.
• Marsaglia, George (2004). “Evaluating the Normal
Distribution”. Journal of Statistical Software. 11
(4). doi:10.18637/jss.v011.i04.
• Maxwell, James Clerk (1860). “V. Illustrations of
the dynamical theory of gases. — Part I: On the
motions and collisions of perfectly elastic spheres”.
Philosophical Magazine, series 4. 19 (124): 19–32.
doi:10.1080/14786446008642818.
• Monahan, J. F. (1985). “Accuracy in random num-
ber generation”. Mathematics of Computation. 45
(172): 559–568. doi:10.1090/S0025-5718-1985-
0804945-X.
• Patel, Jagdish K.; Read, Campbell B. (1996). Hand-
book of the Normal Distribution (2nd ed.). CRC
Press. ISBN 0-8247-9342-0.
• Pearson, Karl (1901). “On Lines and Planes of
Closest Fit to Systems of Points in Space” (PDF).
Philosophical Magazine. 6. 2: 559–572.
24 21 EXTERNAL LINKS

• Pearson, Karl (1905). "'Das Fehlergesetz und • Normal Distribution Video Tutorial Part 1-2 on
seine Verallgemeinerungen durch Fechner und Pear- YouTube
son'. A rejoinder”. Biometrika. 4 (1): 169–212.
doi:10.2307/2331536. JSTOR 2331536. • Normal distribution calculator

• Pearson, Karl (1920). “Notes on the History
of Correlation”. Biometrika. 13 (1): 25–45.
doi:10.1093/biomet/13.1.25. JSTOR 2331722.
• Rohrbasser, Jean-Marc; Véron, Jacques (2003).
“Wilhelm Lexis: The Normal Length of Life as an
Expression of the “Nature of Things"". Population.
58 (3): 303–322. doi:10.3917/pope.303.0303.
• An 8-foot-tall (2.4 m) Probability Machine (named
Sir Francis) comparing stock market returns to the
randomness of the beans dropping through the quin-
cunx pattern. on YouTube Link originating from
Index Fund Advisors

• Stigler, Stephen M. (1978). “Mathematical Statis-

tics in the Early States”. The Annals of Statis-
tics. 6 (2): 239–265. doi:10.1214/aos/1176344123.
JSTOR 2958876.

• Stigler, Stephen M. (1982). “A Modest Pro-

posal: A New Standard for the Normal”.
The American Statistician. 36 (2): 137–138.
doi:10.2307/2684031. JSTOR 2684031.

• Stigler, Stephen M. (1986). The History of Statis-

tics: The Measurement of Uncertainty before 1900.
Harvard University Press. ISBN 0-674-40340-1.

• Stigler, Stephen M. (1999). Statistics on the Table.

Harvard University Press. ISBN 0-674-83601-4.

• Walker, Helen M. (1985). “De Moivre on the Law

of Normal Probability” (PDF). In Smith, David Eu-
gene. A Source Book in Mathematics. Dover. ISBN
0-486-64690-4.

• Wallace, C. S. (1996). “Fast pseudo-random gen-

erators for normal and exponential variates”. ACM
Transactions on Mathematical Software. 22 (1):
119–127. doi:10.1145/225545.225554.

• Weisstein, Eric W. “Normal Distribution”.

MathWorld.

• West, Graeme (2009). “Better Approximations

to Cumulative Normal Functions” (PDF). Wilmott
Magazine: 70–76.

• Zelen, Marvin; Severo, Norman C. (1964).

Probability Functions (chapter 26). Handbook
of mathematical functions with formulas, graphs,
and mathematical tables, by Abramowitz, M.; and
Stegun, I. A.: National Bureau of Standards. New
York, NY: Dover. ISBN 0-486-61272-4.

21 External links
• Hazewinkel, Michiel, ed. (2001), “Normal distribu-
tion”, Encyclopedia of Mathematics, Springer, ISBN
978-1-55608-010-4
 25

22 Text and image sources, contributors, and licenses

22.1 Text
• Normal distribution Source: https://en.wikipedia.org/wiki/Normal_distribution?oldid=741607056 Contributors: Damian Yerrick, Ax-
elBoldt, Lee Daniel Crocker, Bryan Derksen, Zundark, The Anome, Ed Poor, Fnielsen, XJaM, Miguel~enwiki, Graft, Heron, Ryguasu,
Olivier, Ericd, Rbrwr, Edward, Patrick, Michael Hardy, SGBailey, Vinodmp, Dcljr, Tomi, Mdebets, Stevan White, Ronz, Snoyes, Den fjät-
trade ankan~enwiki, Ciphergoth, Cherkash, Charles Matthews, Stan Lioubomoudrov, Jitse Niesen, Br43402, Prumpf, Taxman, Fibonacci,
Zero0000, AaronSw, Wilke, Robbot, Sander123, Robbyjo~enwiki, Benwing, R3m0t, Naddy, Sverdrup, Henrygb, Nilmerg, Ojigiri~enwiki,
Meelar, Alpharigel, Wile E. Heresiarch, Enochlau, Giftlite, Christopher Parham, Markus Krötzsch, J heisenberg, BenFrantzDale, Wwoods,
Will Thimbleby, Jorge Stolfi, Gzornenplatz, Tristanreid, Utcursch, Pgan002, Knutux, Fangz, Noe, Rajasekaran Deepak, MarkSweep,
Piotrus, Gunnar Larsson, Aylex~enwiki, Gauss, Elektron, Karl-Henner, Urhixidur, Fintor, Fledylids, 4pq1injbok, Rich Farmbrough, Gua-
nabot, Jpk, Michael Zimmermann, Paul August, Bender235, Rubicon, Zaslav, MisterSheik, Pjrich, Laurascudder, Aude, Art LaPella, Sole
Soul, O18, Mysteronald, Aydee, Rajah, Darwinek, NickSchweitzer, Jjron, Jumbuck, Jérôme, Gary, Denis.arnaud, 119, Eric Kvaalen,
Ossiemanners, PAR, Burn, Joris Gillis, Rabarberski, Pontus, Brock, Cburnett, Jheald, MIT Trekkie, Forderud, Kay Dekker, Richwales,
Oleg Alexandrov, Joriki, Cgibbard, Mindmatrix, Xover, RHaworth, LOL, David Haslam, SusanLarson, StradivariusTV, Oliphaunt, Ja-
cobolus, Zrenneh, Drostie, Eclecticos, Dzordzm, Yoshis88, Btyner, Graham87, RxS, AllanBz, Rjwilmsi, Seidenstud, Zbxgscqf, Salix alba,
Bubba73, Syced, JanSuchy, Amanjain, Seliopou, Ian Pitchford, RexNL, Dannya222, Fresheneesz, TeaDrinker, ThePlaz, Waagh~enwiki,
Kri, Pcody, Nicholasink, DVdm, YurikBot, Wavelength, Personman, Alektzin, Joebeone, Hydrargyrum, Stephenb, Bruguiea, Trollderella,
Schmock, Cruise, Tony1, Bota47, Smaines, Nikkimaria, Boreas231, Arthur Rubin, Fang Aili, Petri Krohn, JahJah, HereToHelp, Marlas-
dad, Mebden, Habbie, Bo Jacoby, Zvika, Cmglee, Finell, Fjdulles, Edin1, Dudzcom, SmackBot, RDBury, KMcD, Prodego, McGeddon,
Unyoyega, AnOddName, Gilliam, Ohnoitsjamie, Chris the speller, Bluebot, Dycotiles, TimBentley, Sirex98, Nbarth, DHN-bot~enwiki,
Bsilverthorn, Iwaterpolo, Can't sleep, clown will eat me, Chlewbot, Jmlk17, Lansey, Morqueozwald, Wen D House, Cybercobra, Mr
Minchin, G716, Mwtoews, Romanski, Meni Rosenfeld, Bidabadi~enwiki, Yoshigev, Wissons, Heimstern, ML5, Jim.belk, Tryggvi bt, A.
Parrot, Loadmaster, JHunterJ, Beetstra, Hiiiiiiiiiiiiiiiiiiiii, Josephus78, Hu12, Iridescent, Chris53516, Vrkaul, CapitalR, A. Pichler, Gil
Gamesh, Pfeldman, Courcelles, CRGreathouse, Shuhao, Jackzhp, Crescentnebula, Shorespirit, Scohoust, CBM, Ali Obeid, KipKnight,
WeggeBot, Neelix, Mr. okinawa, Lixy, 137 0, FilipeS, AndrewHowse, Dr.enh, P.jansson, Siddiganas, Mikewax, Zaracattle, Saintrain,
Talgalili, Epbr123, Tdunning, Mirams, Headbomb, Mm 202, Leon7, Pablomme, DanSoper, AntiVandalBot, Wainson, Gioto, Widefox,
Paclopes, Seaphoto, Lself, Edokter, Pabristow, Waerloeg, Jason.yosinski, Chill doubt, Gökhan, Trewin, JAnDbot, Plantsurfer, Mcorazao,
Jeff560, Owenozier, Pdumon, Ikanreed, Mkch, Coolhandscot, Tstrobaugh, MSBOT, Coffee2theorems, Magioladitis, Javazen, VoABot II,
A.M.R., SHCarter, Baccyak4H, Sullivan.t.j, A3nm, David Eppstein, User A1, Polyester, JaGa, Mateoee, Doood1, Dima373, Mozza314,
Lilac Soul, Manticore, Jorgenumata, Abecedare, Ginsengbomb, Hughperkins, TheSeven, Troutinthemilk, Kedawa, It Is Me Here, Mikael
Häggström, Coppertwig, Jia.meng, Andraaide, SJP, Policron, Shaww, DavidCBryant, DorganBot, Gemini1980, VolkovBot, DrMicro,
HughD, Smidas3, Dj thegreat, Appoose, Anna Lincoln, LiranKatzir, Seraphim, The Tetrast, Mundhenk, KP-Adhikari, Meters, Joseph A.
Spadaro, Ramiromagalhaes, Markhebner, Why Not A Duck, Junkinbomb, Nolanbard, Scottkosty, Quietbritishjim, SieBot, Ivan Štambuk,
Rlendog, Gex999, Metaxal, Cwkmail, Keilana, Charles Wolf, Cffk, Hxhbot, RSStockdale, OKBot, MrZeebo, Lscharen, Svick, Water
and Land, Josuechan, Kjtobo, Randomblue, Melcombe, DEMcAdams, ClueBot, David.hilton.p, DFRussia, Jpsauro, Superdupereditor,
Plastikspork, EelkeSpaak, Winsteps, MATThematical, Jorisverbiest, GibboEFC, MrKIA11, SamuelTheGhost, Gauravm1312, Awickert,
Tomeasy, Shabbychef, Egorre, Lartoven, Sun Creator, NuclearWarfare, Cenarium, Livius3, Vhlafuente, SergioBruno66, ChrisHodgesUK,
Evan Manning, Vakulgupta, Bluemaster, Qwfp, SoxBot III, Rishi.bedi, DumZiBoT, XLinkBot, Abtweed98, SilvonenBot, Bhockey10,
Jgonion, Paulpeeling, GDibyendu, Addbot, Somebody9973, Mortense, Willking1979, Tcncv, Fgnievinski, Neodop, Vijayarya, Mrocklin,
Lpele, MrOllie, Eivindbot, LaaknorBot, Uscitizenjason, Renatokeshet, Lightbot, Matěj Grabovský, Luckas-bot, Yobot, Qonnec, Rsquire3,
Wjastle, JBancroftBrown, AnomieBOT, Erel Segal, Message From Xenu, Tucoxn, Hiihammuk, Deepakazad, Citation bot, Zhurov, Sergey
Suslov, EOBarnett, GB fan, ArthurBot, Xqbot, Bdmy, Wakamex, Drilnoth, Hugo gasca aragon, Gigemag76, Ptrf, Achalddave, Thekil-
luminati, Almabot, ComputerPsych, Chen-Pan Liao, Kaslanidi, Samwb123, Podgy piglet, FrescoBot, Nicolas Perrault III, Nixphoeni,
Gandrusz, Geekinajeep, Citation bot 1, NoahDawg, DrilBot, Boxplot, Pinethicket, I dream of horses, Elockid, Stpasha, Aby.vinod, Gper-
jim, Dude1818, Firelog, WardMuylaert, TobeBot, Trappist the monk, Fergusq, Jonkerz, Tschwertner, Duoduoduo, Enthdegree, Aurimus,
Stroppolo, RjwilmsiBot, TjBot, Kastchei, Millerdl, Beyond My Ken, GordontheGorgon, EmausBot, 5:40, John of Reading, Scm, Laurifer,
JaeDyWolf, Tommy2010, Netheril96, K6ka, Coubure, Offsure, Chharvey, R. J. Mathar, Derekleungtszhei, JA(000)Davidson, A930913,
Quondum, Galastril, SporkBot, AManWithNoPlan, Arnaugir, Jeffjnet, Supersidvicious, ClueBot NG, Asitgoes, Mathstat, Wikidilworth,
Cheungpuiho04, มือใหม่, Abhijeet Safai, Frietjes, Tianshinihao8, Hikenstuff, Alex.j.flint, Pat5547, Timflutre, Helpful Pixie Bot, BG19bot,
StijnDeVuyst, Jan Spousta, Davidiad, Itsapatel, Duxwing, Statguy1, Ivan trus, Mu.ting, Dexbot, Schbam, AppliedMathematics, SFK2,
Brirush, Limit-theorem, Naereen, Jamesmcmahon0, Nilstec inc, Piotr Jurkiewicz, Jurgengripp, Dr.alaagad, Paul2520, Catalin.ghervase,
314piwm, Math buff, , Tertius51, BeyondNormality, Yura1987, Monkbot, Virion123, Couldbegu, Sotirios-d, Metaprone, JC713,
Velvel2, Edthewikieditor, Jonnyt43, Boky90, Isambard Kingdom, Kiszoltansandor, Darth Meeseus, Lordvig, 5D2262B74, Ashok1mukesh,
Tentonne, Kakutani616, Kernel estimator, StrangeIB, Niuman Comas and Anonymous: 696

22.2 Images
• File:Carl_Friedrich_Gauss.jpg Source: https://upload.wikimedia.org/wikipedia/commons/9/9b/Carl_Friedrich_Gauss.jpg License:
Public domain Contributors: Gauß-Gesellschaft Göttingen e.V. (Foto: A. Wittmann). Original artist: After Christian Albrecht Jensen
• File:Commons-logo.svg Source: https://upload.wikimedia.org/wikipedia/en/4/4a/Commons-logo.svg License: CC-BY-SA-3.0 Contribu-
tors: ? Original artist: ?
• File:De_moivre-laplace.gif Source: https://upload.wikimedia.org/wikipedia/commons/0/06/De_moivre-laplace.gif License: Public do-
main Contributors: Own work Original artist: Stpasha
• File:Dice_sum_central_limit_theorem.svg Source: https://upload.wikimedia.org/wikipedia/commons/8/8c/Dice_sum_central_limit_
theorem.svg License: CC BY-SA 3.0 Contributors: Own work Original artist: Cmglee
• File:Empirical_Rule.PNG Source: https://upload.wikimedia.org/wikipedia/commons/a/a9/Empirical_Rule.PNG License: CC BY-SA
4.0 Contributors: Own work Original artist: Dan Kernler
26 22 TEXT AND IMAGE SOURCES, CONTRIBUTORS, AND LICENSES

• File:Fisher_iris_versicolor_sepalwidth.svg Source: https://upload.wikimedia.org/wikipedia/commons/4/40/Fisher_iris_versicolor_

sepalwidth.svg License: CC BY-SA 3.0 Contributors: en:Image:Fisher iris versicolor sepalwidth.png Original artist: en:User:Qwfp (origi-
nal); Pbroks13 (talk) (redraw)
• File:FitNormDistr.tif Source: https://upload.wikimedia.org/wikipedia/commons/d/d8/FitNormDistr.tif License: Public domain Contrib-
utors: Own work Original artist: Buenas días
• File:Free-to-read_lock_75.svg Source: https://upload.wikimedia.org/wikipedia/commons/8/80/Free-to-read_lock_75.svg License: CC0
Contributors: Adapted from 9px|Open_Access_logo_PLoS_white_green.svg Original artist: This version:Trappist_the_monk (talk)
(Uploads)
• File:Normal_Distribution_CDF.svg Source: https://upload.wikimedia.org/wikipedia/commons/c/ca/Normal_Distribution_CDF.svg
License: Public domain Contributors: self-made, Mathematica, Inkscape Original artist: Inductiveload
• File:Normal_Distribution_PDF.svg Source: https://upload.wikimedia.org/wikipedia/commons/7/74/Normal_Distribution_PDF.svg
License: Public domain Contributors: self-made, Mathematica, Inkscape Original artist: Inductiveload
• File:OEISicon_light.svg Source: https://upload.wikimedia.org/wikipedia/commons/d/d8/OEISicon_light.svg License: Public domain
Contributors: Own work Original artist: Watchduck (a.k.a. Tilman Piesk)
• File:Pierre-Simon_Laplace.jpg Source: https://upload.wikimedia.org/wikipedia/commons/e/e3/Pierre-Simon_Laplace.jpg License:
Public domain Contributors: This image appears identical to the cover image used by Gillispie et al. They cite the portrait as an 1842
posthumous portrait by Madame Feytaud, courtesy of the Académie des Sciences, Paris. Original artist: Sophie Feytaud (fl.1841)
• File:Planche_de_Galton.jpg Source: https://upload.wikimedia.org/wikipedia/commons/2/21/Planche_de_Galton.jpg License: CC BY-
SA 3.0 Contributors: Own work Original artist: Antoine Taveneaux
• File:QHarmonicOscillator.png Source: https://upload.wikimedia.org/wikipedia/commons/b/bb/QHarmonicOscillator.png License:
CC-BY-SA-3.0 Contributors: Own work Original artist: en:User:FlorianMarquardt

22.3 Content license

• Creative Commons Attribution-Share Alike 3.0