Sunteți pe pagina 1din 2

Distribution

Equation

More Information

Power-law

Y = mX^n

Linear equation ;
Lny= n lnx + lnm

Benfords law

Zipfs law
Binomial
distribution

Pareto distributions

Normal
Distritbution
Lognormal
Distritbution
Poission
Distritbution
Exponential
Distritbution
Cumulative
distribution
function [CDF]
Kolmogorov
Smirnoff test
Regression toward
the mean

Pr(d ) ln( d 1) ln( d )


or Pr( d ) log10 (1 d1 )

Pr( X r )

Ar

Pr( x ) nx p 1 p
x

Pr( X r )

n x

Ar

Description
Bigger means more efficient
and n usually divided by 4
The probability of number 1
to occur is 30% as A leading
digit.
A leading digit is the first slot
in a number. Why is that
because it is hard to double
from 1 to 2 but easy to
increase form 9 to 10.
If we have a ranking and we
would like to know what is
the probability something to
happe.. Rank * size = a
constant number
To know the probability of
success of failure. It is
discreet (trails)
It must have categories. We
rank form big value to small .
calculate the percentage of
all. We add the percentages
each one with the previous
one. We draw the grapg of
values and their pareto
percentage. Then we
consider the 80% as the
important to solve
The Majority of the sample
represents 2/3.

Time related

Compare the distance


between the optimal line and
your line

Discrete and Continuous variables;


variables
graph

rule

Discrete
Can be listed and counted
Bar charts

Continuous
Tend to be infinite (like weight)
Continuous line

Area and probability = 1

S-ar putea să vă placă și