Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

%Error in JASP vs Bayes Error Rate

Hi,

I can't get my head around how "% error" in JASP possibly relates to the Bayes Error Rate (BER). It seems that both types of measure relates to measuring the Irreducible Error in data. But from my understanding the "% error" is a proportional error in the sense that, for example, a 5% error of a BF of 3 means that the BF is with certainty (unclear with what certainty) somewhere between 2.85 and 3.15? Meanwhile, I understand the BER such as the proportion of times that a model lead to wrongful predictions, so that a 5% error would mean that the BF.

Hope I made enough sense to start a discussion...

Best regards,

Philip

Comments

  • Hi Philip,

    The %error gives an indication of how accurately the numerical methods have estimated the BF. I think that it's the standard error expressed as a proportion of the BF, but the BayesFactor package should explain what it is exactly. When in doubt, you can email Richard Morey (and please let us know what he says!)

    BER is a different concept that relates to the performance of the method in repeated use (it seems).

    Cheers,
    E.J.

Sign In or Register to comment.