skewness e kurtosis

Duas medidas importantes para caracterizar uma distribuição não-normal são os coeficientesde skewness e de
kurtosis.

- No caso do skewness, coeficiente próximo de zero significa simetria, caso contrário, uma tendência à
esquerda para números negativos e, à direita para números positivos.

-A kurtosis mede a concentração próxima a média (ou pico),a Kurtosis (Curtose) é uma medida de dispersão
que caracteriza o "achatamento" da curva da função de distribuição.
Se o valor da Kurtosis for = 0, então tem o mesmo achatamento que a distribuição normal.
Se o valor é > 0 então a distribuição em questão é mais alta (afunilada) e concentrada que a distribuição
normal.
Se o valor é < 0 então a função de distribuição é mais "achatada" que a distribuição normal

The Shapiro-Wilk Test For Normality

The Shapiro-Wilk test, proposed in 1965, calculates a W statistic that tests whether a random sample, x1, x2, ...,
xn comes from (specifically) a normal distribution . Small values of W are evidence of departure from normality
and percentage points for the W statistic, obtained via Monte Carlo simulations, were reproduced by Pearson
and Hartley (1972, Table 16). This test has done very well in comparison studies with other goodness of fit
tests.

The W statistic is calculated as follows:

where the x(i) are the ordered sample values (x(1) is the smallest) and the ai are constants generated from the
means, variances and covariances of the order statistics of a sample of size n from a normal distribution (see
Pearson and Hartley (1972, Table 15).

For more information about the Shapiro-Wilk test the reader is referred to the original Shapiro and Wilk (1965)
paper and the tables in Pearson and Hartley (1972),

Chi-Square Goodness-of-Fit Test

Purpose: The chi-square test (Snedecor and Cochran, 1989) is used to test if a sample of data came from a
Test for population with a specific distribution.
distributional
adequacy An attractive feature of the chi-square goodness-of-fit test is that it can be applied to any univariate
distribution for which you can calculate the cumulative distribution function. The chi-square goodness-of-
fit test is applied to binned data (i.e., data put into classes). This is actually not a restriction since for
non-binned data you can simply calculate a histogram or frequency table before generating the chi-
square test. However, the value of the chi-square test statistic are dependent on how the data is binned.
Another disadvantage of the chi-square test is that it requires a sufficient sample size in order for the
chi-square approximation to be valid.

The chi-square test is an alternative to the Anderson-Darling and Kolmogorov-Smirnov goodness-of-fit

tests. The chi-square goodness-of-fit test can be applied to discrete distributions such as the binomial
and the Poisson. The Kolmogorov-Smirnov and Anderson-Darling tests are restricted to continuous
distributions.

Additional discussion of the chi-square goodness-of-fit test is contained in the product and process
comparisons chapter (chapter 7).
Definition The chi-square test is defined for the hypothesis:
H0: The data follow a specified distribution.
 Ha: The data do not follow the specified distribution.
Jarque–Bera
Test Statistic: test
For the chi-square goodness-of-fit computation, the data are divided into k bins and
In statistics, the
theJarque–Bera test
test statistic is is a goodness-of-fit
defined as measure of departure from
normality, based on the sample kurtosis and skewness. The test is named after Carlos
Jarque and Anil K. Bera. The test statistic JB is defined as

where is the observed frequency for bin i and is the expected frequency for bin
i. The expected frequency is calculated by
where n is the number of observations (or degrees of freedom in general); S is the sample
skewness, and K is the sample kurtosis:
where F is the cumulative Distribution function for the distribution being tested, Yu is
the upper limit for class i, Yl is the lower limit for class i, and N is the sample size.

This test is sensitive to the choice of bins. There is no optimal choice for the bin
width (since the optimal bin width depends on the distribution). Most reasonable
choices should produce similar, but not identical, results. Dataplot uses 0.3*s, where
s is the sample standard deviation, for the class width. The lower and upper bins are
at the sample mean plus and minus 6.0*s, respectively. For the chi-square
approximation to be valid, the expected frequency should be at least 5. This test is
not valid for small samples, and if some of the counts are less than five, you may
where and need aretothe estimates
combine some of bins
third in
and
thefourth
tails. central moments, respectively, is
the sample
Significance mean,. and is the estimate of the second central moment, the variance.
Level:
The statistic JBThe
Critical hastestan statistic
asymptotic chi-square
follows, distribution
approximately, with two degrees
a chi-square of freedom
distribution with (k and
- c) degrees
can be used tooftest
Region: the nullwhere
freedom hypothesis
k is thethat the data
number are from acells
of non-empty normalanddistribution.
c = the numberThe null
of
hypothesis is aestimated
joint hypothesis of the(including
parameters skewnesslocation
being zeroand and the
scale excess kurtosis
parameters being
and shape
0, since samples from a normal
parameters) distribution
for the distributionhave
+ 1.an expected
For example, skewness of 0 and an
for a 3-parameter Weibull
expected excess kurtosis of
distribution, c = 4.0 (which is the same as a kurtosis of 3). As the definition of JB
shows, any deviation from this increases the JB statistic.
Therefore, the hypothesis that the data are from a population with the specified
The chi-squaredistribution
approximation, however,
is rejected if is overly sensitive (lacking specificity) for small
samples, rejecting the null hypothesis often when it is in fact true. Furthermore, the
distribution of p-values departs from a uniform distribution and becomes a right-skewed uni-
modal distribution, especially for small p-values. This leads to a large Type I error rate. The
table below shows some p-values approximated by a chi-square distribution
with k -that differ from
where is the chi-square percent point function c degrees of freedom
their true alphaandlevels for very small samples.
a significance level of .

Calculated
In thep-value equivalents
above formulas for to
thetrue alpha
critical levels the
regions, at given sample
Handbook sizes the convention
follows
True α level 20 30 50 70 100
that.1 is the upper
.307critical.252
value from the chi-square
.201 .183 distribution
.1560 and is the
lower critical value from the chi-square distribution. Note that this is the opposite of
.05 .1461 .109 .079 .067 .062
what is used in some texts and software programs. In particular, Dataplot uses the
.025 convention.
opposite .051 .0303 .020 .016 .0168
.01 .0064 .0033 .0015 .0012 .002
(These values have been approximated by using Monte Carlo simulation on Matlab)

Purpose: As seen in MATLAB, the chi-square approximation for the JB statistic's distribution is only
Test for used for large sample sizes (> 2000). For smaller sample sizes, it uses a table derived from
Monte Carlo simulations in order to interpolate p-values for smaller samples. [1]

Kolmogorov-Smirnov Goodness-of-Fit Test

Distributional Adequacy
The Kolmogorov-Smirnov test (Chakravart, Laha, and Roy, 1967) is used to decide if a sample comes from a
population with a specific distribution.
The Kolmogorov-Smirnov (K-S) test is based on the empirical distribution function (ECDF). Given N ordered data
points Y1, Y2, ..., YN, the ECDF is defined as

where n(i) is the number of points less than Yi and the Yi are ordered from smallest to largest value. This is a step
function that increases by 1/N at the value of each ordered data point.

The graph below is a plot of the empirical distribution function with a normal cumulative distribution function for 100
normal random numbers. The K-S test is based on the maximum distance between these two curves.

Characteristics and Limitations of the K-S Test

An attractive feature of this test is that the distribution of the K-S test statistic itself does not depend on the underlying
cumulative distribution function being tested. Another advantage is that it is an exact test (the chi-square goodness-of-
fit test depends on an adequate sample size for the approximations to be valid). Despite these advantages, the K-S
test has several important limitations:

1. It only applies to continuous distributions.

2. It tends to be more sensitive near the center of the distribution than at the tails.
3. Perhaps the most serious limitation is that the distribution must be fully specified. That is, if location, scale,
and shape parameters are estimated from the data, the critical region of the K-S test is no longer valid. It
typically must be determined by simulation.

Due to limitations 2 and 3 above, many analysts prefer to use the Anderson-Darling goodness-of-fit test. However, the
Anderson-Darling test is only available for a few specific distributions.
Definition
The Kolmogorov-Smirnov test is defined by:
H0: The data follow a specified distribution
Ha: The data do not follow the specified distribution
Test The Kolmogorov-Smirnov test statistic is defined as
Statistic:

where F is the theoretical cumulative distribution of the

distribution being tested which must be a continuous
distribution (i.e., no discrete distributions such as the
binomial or Poisson), and it must be fully specified (i.e., the
location, scale, and shape parameters cannot be estimated
from the data).
Significance .
Level:
Critical The hypothesis regarding the distributional form is rejected
Values: if the test statistic, D, is greater than the critical value
obtained from a table. There are several variations of these
tables in the literature that use somewhat different scalings
for the K-S test statistic and critical regions. These
alternative formulations should be equivalent, but it is
 necessary to ensure that the test statistic is calculated in a
way that is consistent with how the critical values were
tabulated.

We do not provide the K-S tables in the Handbook since

software programs that perform a K-S test will provide the
relevant critical values.

Technical Note
Previous editions of e-Handbook gave the following formula for the computation of the Kolmogorov-Smirnov goodness
of fit statistic:

This formula is in fact not correct. Note that this formula can be rewritten as:

This form makes it clear that an upper bound on the difference between these two formulas is i/N. For actual data, the
difference is likely to be less than the upper bound.

For example, for N = 20, the upper bound on the difference between these two formulas is 0.05 (for comparison, the
5% critical value is 0.294). For N = 100, the upper bound is 0.001. In practice, if you have moderate to large sample
sizes (say N ≥ 50), these formulas are essentially equivalent.

Anderson-Darling
Test
Purpose: The Anderson-Darling test (Stephens, 1974) is used to test if a sample of data
Test for came from a population with a specific distribution. It is a modification of the
Distributional Kolmogorov-Smirnov (K-S) test and gives more weight to the tails than does
Adequacy the K-S test. The K-S test is distribution free in the sense that the critical values
do not depend on the specific distribution being tested. The Anderson-Darling
test makes use of the specific distribution in calculating critical values. This has
the advantage of allowing a more sensitive test and the disadvantage that
critical values must be calculated for each distribution. Currently, tables of
critical values are available for the normal, lognormal, exponential, Weibull,
extreme value type I, and logistic distributions. We do not provide the tables of
critical values in this Handbook (see Stephens 1974, 1976, 1977, and 1979)
since this test is usually applied with a statistical software program that will
print the relevant critical values.

The Anderson-Darling test is an alternative to the chi-square and Kolmogorov-

Smirnov goodness-of-fit tests.
Definition The Anderson-Darling test is defined as:
H0: The data follow a specified distribution.
Ha: The data do not follow the specified distribution
Test The Anderson-Darling test statistic is defined as
Statistic:
where

F is the cumulative distribution function of the specified

distribution. Note that the Yi are the ordered data.
Significance
Level:
Critical The critical values for the Anderson-Darling test are
Region: dependent on the specific distribution that is being tested.
Tabulated values and formulas have been published
 (Stephens, 1974, 1976, 1977, 1979) for a few specific
distributions (normal, lognormal, exponential, Weibull,
logistic, extreme value type 1). The test is a one-sided test
and the hypothesis that the distribution is of a specific form
is rejected if the test statistic, A, is greater than the critical
value.

Note that for a given distribution, the Anderson-Darling

statistic may be multiplied by a constant (which usually
depends on the sample size, n). These constants are given
in the various papers by Stephens. In the sample output
below, this is the "adjusted Anderson-Darling" statistic. This
is what should be compared against the critical values. Also,
be aware that different constants (and therefore critical
values) have been published. You just need to be aware of
what constant was used for a given set of critical values (the
needed constant is typically given with the critical values).

Lilliefors

The test proceeds as follows:

1. First estimate the population mean and population variance based on the data.
2. Then find the maximum discrepancy between the empirical distribution function and the cumulative distribution
function (CDF) of the normal distribution with the estimated mean and estimated variance. Just as in the Kolmogorov–
Smirnov test, this will be the test statistic.
3. Finally, we confront the question of whether the maximum discrepancy is large enough to be statistically significant,
thus requiring rejection of the null hypothesis. This is where this test becomes more complicated than the
Kolmogorov–Smirnov test. Since the hypothesized CDF has been moved closer to the data by estimation based on
those data, the maximum discrepancy has been made smaller than it would have been if the null hypothesis had
singled out just one normal distribution. Thus the "null distribution" of the test statistic, i.e. its probability distribution
assuming the null hypothesis is true, is stochastically smaller than the Kolmogorov–Smirnov distribution. This is the
Lilliefors distribution.