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

A Number of pseudoexperiments

  1. Jul 6, 2016 #1


    User Avatar
    Gold Member

    How can someone decide what's the best number of produced pseudoexperiments in his set up?
    In particular I have a value [itex]N[/itex] which I want to vary wrt 2 nuisance parameters with relative uncertainties [itex]\delta_1,\delta_2[/itex]. I am producing [itex]n[/itex] pseudoexperiments (samples) in each calculating the mean and the standard deviation of
    [itex]N_i =N_i^0 \Big[1 + \delta_1 \mathcal{N}_1(0,1) + \delta_2 \mathcal{N}_2(0,1) \Big][/itex]
    How can I decide whether the [itex]n[/itex]-trials I am choosing is optimal?
    I have reached the following conclusion after some thinking but I am not sure.... the sample's relative uncertainty that is [itex]\sqrt{\text{Var}(N)}/\bar{N}[/itex] should be as small as possible... is that a correct way?
  2. jcsd
  3. Jul 6, 2016 #2

    Vanadium 50

    User Avatar
    Staff Emeritus
    Science Advisor
    Education Advisor
    2017 Award

    The answer is "as many as you can". There is no point where another pseudo-experiment makes things worse and not better.
  4. Jul 6, 2016 #3


    User Avatar
    Gold Member

    In general I've seen plots where they show the "observed" let's say value (that they vary in each PE), and the sampled values for different/increasing [itex]n[/itex]-trials (eg 100,1000,10000,1000000)... so, the sampled values get closer to the observed value, but also they get more statistically constrained... I was wondering if by looking at something like this allows someone to decide with what [itex]n[/itex] they should go.
  5. Jul 6, 2016 #4


    User Avatar
    2017 Award

    Staff: Mentor

    As much as your computers can reasonably compute. The uncertainty from the limited number of pseudoexperiments goes down, which is great.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted