Marcello comments on Conditional Independence, and Naive Bayes - Less Wrong

30 Post author: Eliezer_Yudkowsky 01 March 2008 01:59AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (15)

Sort By: Old

You are viewing a single comment's thread.

Comment author: Marcello 01 March 2008 03:01:12AM 7 points [-]

Nice post. However:

Please /please/ don't use the "+" sign like that! H(X+Y+Z) should be H(X,Y,Z). "So for the joint distribution X+Y there are" should be "So for the joint distribution of X and Y there are" etc. I was skimming your post, misunderstood your meaning entirely and started wondering if you had made a mistake until I went back and noticed that some of your "+"s meant "X and Y" rather than "the value of X plus the value of Y". (So for example, when I read """ Z has two states, "even" and "odd", perfectly correlated to the evenness or oddness of X+Y. In fact, we'll suppose that Z is just the question "Are X+Y even or odd?" """ I thought "golly, 'Are' X+Y even or odd? Must be a grammar mistake.") Now granted, it was easy to tell what you really meant after I slowly read through the introductory paragraph, but please change it, because: * It will confuse mathematicians who are skimming because they've seen things like this before * You use the other sense of "+" in your equations, and there's no excuse for operator overloading if it can be avoided. * The books I've read just say "XY" or "X,Y" if they need a name for the joint variable. * It's not just non-standard, but used inconsistantly. You say H(X+Y+Z) in one place and H(X,Y) in another.