@Dan You must be a quick learner. Somehow I had the impression you had been on the site for over a year.

Heh. This one is easy enough to do in one's head (I did it just now and got (3*19+10)/1000=0.067): stats.stackexchange.com/questions/112561/… ... unfortunately, I don't think the way I did it would be understood by the OP.

I'll see if I can think of a way explain it simply, without making it too much longer.

Hi, I'm a newbie in stats/data-science

could you help me on this one:

I have several descriptions, and I'd like to classify them. I have also a bag-of-words where some words are highly related to one category

Is Naive Bayes a good candidate to do this ?

And should I ask this question on the site ?

@younes0 The first thing to do is to search the site, using the various keywords in your question (and any others you think of) in a few different combinations. When you turn up posts something like what you want, search down the 'Related' column in the sidebar for any others. You can also search the site via google (using the site: argument) - sometimes that may find things the on-site search doesn't - though the on site one is pretty good.

If that doesn't turn up anything that answers your question, then you should post your question on the site.

Thanks @Glen_b, I have found several informations on the Data Scientist stack exchange website which targets Machine Learning enthousiasts like me among others. Great advices altough, I found myself too that googling Stack Exchange content was more efficient than the onsite search.