« first day (2647 days earlier)      last day (2656 days later) » 

07:36
Howdy
 
11 hours later…
19:05
W/i the SE system, we are 8th in terms of Qs / day (#10 by traffic). Of the 7 sites that outrank us, 4 are SO. I also think it's a bit weird that the #3 site is SO in Russian.
 
3 hours later…
22:05
Is there any specific difference between a "distribution" and a "family" of distributions? Do we have to say that N(mu,sigma^2) is a family because it's indexed by two parameters? Or can one say that it's a distribution indexed by two params? Wikipedia says things like "gamma distribution is a two-parameter family...", but at the same time "binomial distribution with parameters n and p is the discrete probability distribution..." so it's not very consistent.
Intuitively I would say that if it's is indexed by a finite set of numbers then it's a distribution with parameters, and if it's indexed by infinite-dimensional set of functions (like exponential family) then it's a "family". But that's just what I made up, not something I've ever seen defined.
Anyone here is familiar with the Cox PH survival model? I'm trying make my own implementation in Tensorflow as an exercise
amoeba, a specific distribution is called a distribution when the parameters have a specified values
A family of distribution is pointing to the general set of possible distribution assuming a certain form

« first day (2647 days earlier)      last day (2656 days later) »