A Randomized Central Limit Theorem

Iddo Eliazar, Joseph Klafter

Research output: Contribution to journalArticlepeer-review

Abstract

The Central Limit Theorem (CLT), one of the most elemental pillars of Probability Theory and Statistical Physics, asserts that: the universal probability law of large aggregates of independent and identically distributed random summands with zero mean and finite variance, scaled by the square root of the aggregate-size (sqrt(n)), is Gaussian. The scaling scheme of the CLT is deterministic and uniform - scaling all aggregate-summands by the common and deterministic factor sqrt(n). This Letter considers scaling schemes which are stochastic and non-uniform, and presents a "Randomized Central Limit Theorem" (RCLT): we establish a class of random scaling schemes which yields universal probability laws of large aggregates of independent and identically distributed random summands. The RCLT universal probability laws, in turn, are the one-sided and the symmetric Lévy laws.

Original languageEnglish
Pages (from-to)290-293
Number of pages4
JournalChemical Physics
Volume370
Issue number1-3
DOIs
StatePublished - 12 May 2010

Keywords

  • Central Limit Theorem (CLT)
  • One-sided Lévy laws
  • Poisson processes
  • Power-laws
  • Randomized Central Limit Theorem (RCLT)
  • Symmetric Lévy laws
  • Universality

Fingerprint

Dive into the research topics of 'A Randomized Central Limit Theorem'. Together they form a unique fingerprint.

Cite this