On Fri, 6 Apr 2001, Satoshi Ikarashi wrote:
> ST> Or extend that example. You have requests at 1am, 5am, 9am, 1pm. x = 4,
> ST> m = 12 hours, and analog calculates that the rate is
> ST> (x - 1) / m
> ST> = 3 / (12 hours)
> ST> = 1 / (4 hours).
> ST> Your formula, x / m, would give 1 / (3 hours) which is wrong.
> ST>
> ST> I don't see how I can explain it any better than this.
>
> Finally, I believe I understand what you mean.
>
> If I have requests at 1am, 5am, 9am, 1pm during time period 0:00
> to 24:00, x = 4 --- the number of requests is 4. On the other
> hand, for the average calculation, you assume the first request
> had come at 0:00 and the last (the 4th) request had come at
> 24:00 (which is 0:00 of the next day). So the rate is (4-1) per
> day.
>
> Is my comprehension correct?
Nearly, but not quite. I assume that the number of requests is 4 _in 12
hours_. So the rate is
(4 - 1) / (12 hours)
= 1 / (4 hours).
If we knew that these were all the requests in the period 00:00 to 23:59,
then we would indeed use your calculation
4 / (24 hours)
= 1 / (6 hours)
But there is no way to know the period covered by the logfile except when
there are requests. Analog can't know that the logfile covers the period
00:00 to 23:59 because there is only evidence that it covers 01:00 to 13:00.
--
Stephen Turner http://www.statslab.cam.ac.uk/~sret1/
Statistical Laboratory, Wilberforce Road, Cambridge, CB3 0WB, England
"Your account can only be used for a single internet session at any one
time and for no more than 24 hours in any one day." (NTL terms of use)
+------------------------------------------------------------------------
| This is the analog-help mailing list. To unsubscribe from this
| mailing list, go to
| http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
| List archives are available at
| http://www.mail-archive.com/[email protected]/
| http://lists.isite.net/listgate/analog-help/archives/
+------------------------------------------------------------------------