1
$\begingroup$

How would you derive marginalization as it is given here: http://en.wikipedia.org/wiki/Marginal_distribution?

Thanks,

$\endgroup$
  • 1
    $\begingroup$ It is consensus that questions be self contained, so please write whatever it is that you're referencing in the page in your question's body. Thank you! $\endgroup$ – Pedro Tamaroff Jun 16 '12 at 2:49
  • 2
    $\begingroup$ The Wikipedia article you refer to has a good deal of detail, with some worked examples. If you have a concrete problem you are having difficulty with, perhaps you could give it explicitly. $\endgroup$ – André Nicolas Jun 16 '12 at 3:25
  • $\begingroup$ @PeterTamaroff The wikipedia page gives the PDF and says to marginalize you integrate. How would you derive marginalization starting from I assume Bayes Theorem? $\endgroup$ – Eiyrioü von Kauyf Jun 18 '12 at 18:40
  • $\begingroup$ A very good question. Perhaps you get better answers if you would state it as: "How to prove the Chapman–Kolmogorov equation?" (in its general form, not the variant limited to Markov chains) $\endgroup$ – Anne van Rossum Dec 21 '14 at 12:33
2
$\begingroup$

Marginalizing amounts to integrating or summing out from joint distributions of several variables to the distribution of a smaller number of variables, usually to one variable. So for example if X and Y have the joint desnsity f(x,y) to get the maginal density for X you compute g(x)=∫f(x,y)dy integrated over all possible values of y. You do the analogous thing with sums for discrete variables.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.