[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Statistics on non-free usage

On 2004-01-09 16:44:34 +0000 Sven Luther <sven.luther@wanadoo.fr> wrote:

The error rate is the proportion misclassified, which cannot be infinite under any circumstances where there are some subjects. I'm also not sure how it applies here.
Ah, yes, that is the stuff taken in the other direction, but then, the
error ratio in my example would be 0 (as 0 correct over 5 total).
speaking of 0 error ratio doesn't seem reasonable. Maybe 100 % error
ratio would be what we want here.

No, the error rate would be 1 = 100% if we misclassified everyone, which is what you seemed to claim. My definition is from B S Everitt's Dictionary of Statistics.

Anyway, it is clear that the data provided by John were crap, and cannot
be used.

Actually, John's tables didn't show that error rate, but they did wash the results you're interested in away in the rounding.

There is probably a non-zero error, but I think it would need to be greater than .5% before you can claim the table is showing the wrong number.
It shows 0, so if at least one person uses it (like me) then we already
have 100% wrong values, no ?

No, they are %.

MJR/slef     My Opinion Only and possibly not of any group I know.
Please http://remember.to/edit_messages on lists to be sure I read
http://mjr.towers.org.uk/ gopher://g.towers.org.uk/ slef@jabber.at
 Creative copyleft computing services via http://www.ttllp.co.uk/

Reply to: