Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Classical definition of probability & kolmogorovs axioms

  1. Feb 11, 2013 #1
    I've seen in some probability theory books that the classical definition of probability is a probability measure, it seems fairly trivial but what is the proof for this? Wikipedia gives a very brief one using cardinality of sets. Is there any other way?
     
  2. jcsd
  3. Feb 11, 2013 #2

    micromass

    User Avatar
    Staff Emeritus
    Science Advisor
    Education Advisor
    2016 Award

    There is no proof. It's a definition. Definitions never have proofs. The only thing you can ask is why this definition is the right one and why it encapsulates our naive understanding of probability. To answer this question, you need to work out some examples of probability spaces. For example, take two dice and throw them. Describe the probability space and see if the theory matches the experiment.
     
  4. Feb 13, 2013 #3

    ssd

    User Avatar

    One is Kolmogorov's axiomatic approach, as you mentioned in the title. Another is Baye's conditional approach...which is a different measure than the classical one.
     
  5. Feb 13, 2013 #4
    oh okay, so you prove that it obeys the axioms by kolmogorov?
     
  6. Feb 14, 2013 #5

    ssd

    User Avatar

    Which one? Classical defn.? The answer is no, these two are different measures.
     
  7. Feb 14, 2013 #6

    Using fairly basic set theory principals and Kolmogorov's 3 basic axioms you can prove/create more postulates (much like the Euclidian approach to geometry). Then from there you can show how and why things in classical probability work the way they do. You can also state Bayes laws in terms of Kolomogorov axioms.

    -Dave K
     
  8. Feb 14, 2013 #7

    Stephen Tashi

    User Avatar
    Science Advisor

    I doubt it. As micromass said, definitions don't have proofs. A definition is not a theorem. Perhaps what you saw in the Wikipedia was a demonstration that the cardinality function on finite sets statisfies the definition of a measure. Such a demonstration is an example of a measure, not a proof of the definition of a measure. (When a mathematical definition for something is invented it is reassuring to demonstrate that at least one example of the thing exists. Such a demonstration does not make the definition "true". It only indicates that it is not futile to study the thing that is defined.)
     
  9. Feb 14, 2013 #8

    D H

    User Avatar
    Staff Emeritus
    Science Advisor

    Bayes' (not Baye's) theorem is consistent with Kolmogorov's axiomatization of probability. It derives from Kolmogorov's axioms.

    Yes, there is a dispute between frequentists and Bayesianists over the meaning and validity of prior probabilities, but that's a debate over interpretations of probability, not over fundamentals.
     
  10. Feb 14, 2013 #9

    ssd

    User Avatar

    Right. Thanks.

    What I really intended to mean is, through Bayes' approach we (first) had the taste of having a different probability measure than the classical approach. More than a century later we again had another probability measure due to Kolmogorov, differing from the classical one. Historically, Bayes' stood alone for many years fighting(!) with the classical.
     
    Last edited: Feb 14, 2013
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Classical definition of probability & kolmogorovs axioms
  1. Axioms of probability (Replies: 13)

  2. Probability Axioms (Replies: 7)

Loading...