Symmetric set: Difference between revisions
en>Dfqppp formula error fixed |
en>Mgkrupa →References: Added {{Functional Analysis}} footer |
||
Line 1: | Line 1: | ||
The '''order in probability''' notation is used in [[probability theory]] and [[statistical theory]] in direct parallel to the [[big-O notation]] that is standard in [[mathematics]]. Where the [[big-O notation]] deals with the convergence of sequences or sets of ordinary numbers, the order in probability notation deals with [[Convergence of random variables|convergence of sets of random variables]], where convergence is in the sense of [[convergence in probability]].<ref>Dodge, Y. (2003) ''The Oxford Dictionary of Statistical Terms'', OUP. ISBN 0-19-920613-9</ref> | |||
For a set of random variables ''X<sub>n</sub>'' and a corresponding set of constants ''a<sub>n</sub>'' (both indexed by ''n'', which need not be discrete), the notation | |||
:<math>X_n=o_p(a_n) \,</math> | |||
means that the set of values ''X<sub>n</sub>''/''a<sub>n</sub>'' converges to zero in probability as ''n'' approaches an appropriate limit. | |||
Equivalently, ''X''<sub>''n''</sub> = o<sub>''p''</sub>(''a''<sub>''n''</sub>) can be written as ''a''<sub>''n''</sub> o<sub>''p''</sub>(1) where ''X''<sub>''n''</sub> = o<sub>''p''</sub>(1) is defined as, | |||
:<math>\lim_{n \to \infty} P(|X_n| \geq \varepsilon) = 0,</math> | |||
for every positive ε.<ref>Yvonne M. Bishop, Stephen E. Fienberg, Paul W. Holland. (1975,2007) ''Discrete multivariate analysis'', Springer. ISBN 0-387-72805-8, ISBN 978-0-387-72805-6</ref> | |||
The notation, | |||
:<math>X_n=O_p(a_n), \,</math> | |||
means that the set of values ''X<sub>n</sub>''/''a<sub>n</sub>'' is stochastically bounded. That is, for any ε > 0, there exists a finite M > 0 such that, | |||
:<math>P(|X_n/a_n| > M) < \varepsilon,\ \forall n.</math> | |||
<!-- | |||
Let ''H''(''n'') be a [[function (mathematics)]] of ''n'', and ''f''(''X'') be a [[real-valued function]] of a [[random variable]] ''X'', where ''X'' has a [[sample size]] of ''n'' observations. Then [[limit]] as ''n'' goes to [[infinity]] of the [[probability]] that ''f''(''X'')/''H''(''n'') is below an [[arbitrary]] [[positive]] [[real number]] is [[zero]]. That is, | |||
:<math>\forall \varepsilon >0, \lim_{n\rightarrow\infty}\mathbb{P}\left(\left|f(X)/H(n)\right|\geq\varepsilon\right)=0.</math> | |||
--> | |||
==Example== | |||
If <math>(X_n)</math> is a stochastic sequence such that each element has finite variance, then | |||
:<math>X_n - E(X_n) = O_p(\sqrt{\operatorname{var}(X_n)}) \,</math> | |||
(see Theorem 14.4-1 in Bishop et al.) | |||
If, moreover, <math>a_n^{-2}\operatorname{var}(X_n) = \operatorname{var}(a_n^{-1}X_n)</math> is a null sequence for a sequence <math>(a_n)</math> of real numbers, then <math>a_n^{-1}(X_n - E(X_n))</math> converges to zero in probability by [[Chebyshev's inequality]], so | |||
:<math>X_n - E(X_n) = o_p(a_n)</math>. | |||
==References== | |||
<references/> | |||
{{DEFAULTSORT:Big O In Probability Notation}} | |||
[[Category:Mathematical notation]] | |||
[[Category:Probability theory]] | |||
[[Category:Statistical terminology]] |
Latest revision as of 04:20, 10 January 2014
The order in probability notation is used in probability theory and statistical theory in direct parallel to the big-O notation that is standard in mathematics. Where the big-O notation deals with the convergence of sequences or sets of ordinary numbers, the order in probability notation deals with convergence of sets of random variables, where convergence is in the sense of convergence in probability.[1]
For a set of random variables Xn and a corresponding set of constants an (both indexed by n, which need not be discrete), the notation
means that the set of values Xn/an converges to zero in probability as n approaches an appropriate limit. Equivalently, Xn = op(an) can be written as an op(1) where Xn = op(1) is defined as,
for every positive ε.[2]
The notation,
means that the set of values Xn/an is stochastically bounded. That is, for any ε > 0, there exists a finite M > 0 such that,
Example
If is a stochastic sequence such that each element has finite variance, then
(see Theorem 14.4-1 in Bishop et al.)
If, moreover, is a null sequence for a sequence of real numbers, then converges to zero in probability by Chebyshev's inequality, so