[rrd-users] Using the Holt-Winters algorithm for prediction of future values

Faraaz Sareshwala fsareshwala at quantcast.com
Wed Sep 2 20:14:58 CEST 2009

Yeah, I've been searching on the Internet for some time now on how to predict future values based on past trends using RRDtool for some time now but haven't gotten anywhere. I'm not so concerned about that graph in particular, but would like to bring the prediction of future data points to our graphs. Are there any pages out there that describe how to do this?


On Wed, Sep 02, 2009 at 10:45:45AM -0700, Faraaz Sareshwala wrote:
> Thanks for the response! I'm wondering what this term prediction means
> then. I took it as prediction of future data points based on previous
> data points and trends. I've also read that prediction, as it is used
> in RRDtool, could also mean prediction of possible issues based on
> current trends.

It can.  But the graphs you referred to are not using the Holt-Winters
features.  They're simple least-squares linear fit lines.

Do you have a question about using one or the other of these?

rrd-users mailing list
rrd-users at lists.oetiker.ch

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20090902/d1448a52/attachment-0001.htm 

More information about the rrd-users mailing list