<html><body><p><br></p><p>Hi all,</p><p><br></p><p>In my experience, for forecasting it is necessary to include some </p><p>"intelligence" about the nature of data. The "blind" forecasting didn't</p><p>provide me satisfactory results. For example, in my pro-active system </p><p>monitoring Multi-ISP operation (in 2001) alerting and predicting problems</p><p>I had to include calendar, working days and public/school holidays etc., </p><p>and then the system was able to disclose problems coming successfully.</p><p><br></p><p>I see the forecasting as quite strongly field-dependent matter.</p><p><br></p><p>Leos</p><p><br></p><p><br>---------- Původní zpráva ----------<br>Od: Tobias Oetiker <tobi@oetiker.ch><br>Datum: 13. 5. 2012<br>Předmět: Re: [rrd-users] Experiance in HW forcast</p><blockquote>Hi Michael,<br><br>Today Michael Schwartzkopff wrote:<br><br>> Hi,<br>><br>> In the last week I read into the HW forecasting and the implementation in the<br>> rrdtools.<br>><br>> After reading a lot there are still some questions open:<br>><br>> 1) How do I choose the initial values for the fitting parameters alpha, beta<br>> and gamma?<br>><br>> in my experiments alpha behaved nicely and it was no problem to find the<br>> optimal alpha. But the least square function of gamma in the range of ]0,1[<br>> seems to be fractal. So it will be extremely difficult to optimize the<br>> parameters.<br>> Anybody else with this experience?<br>><br>> In all samples in the net nobody explains why he choose exactly this<br>> parameters. Why?<br><br>take some high resolution sample data worth several seasons (weeks)<br>and then try out the different parameters ... note some are very<br>sensitive ... the best way to get your config right is to tune with<br>real data ...<br><br>> 2) Is there any possibility to optimize the fitting parameters to recent data?<br>> My idea would be to optimize the fitting parameters, create the RRDs and to<br>> optimize the RRDs from time to time. Any ideas or comments?<br><br>you can rrdtool tune the parameters ...<br><br><br>> Are there any experts on the list? I would like to start a discussion about<br>> making the use of the forecasting in RRDs as simple as creating a plain graph.<br><br>there is very little discussion on this toppic, but maybe this will<br>start it ...<br><br>cheers<br>tobi<br><br><br>> Greetings,<br>><br>><br><br>-- <br>Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland<br><a href="http://it.oetiker.ch">http://it.oetiker.ch</a> tobi@oetiker.ch ++41 62 775 9902 / sb: -9900<br><br>_______________________________________________<br>rrd-users mailing list<br>rrd-users@lists.oetiker.ch<br><a href="https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users">https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users</a></blockquote></body></html>