[rrd-developers] Instability in HWPREDICT ?

[Stanley Hopcroft]

Dear Ladies and Gentlemen,

I am looking forward to using the HWPREDICT CFs contributed by Mr 
Brutlag.

However, when I do so using

alpha: 0.0035
 beta: 0.1

The DEVPREDICT CF tracks beautifully the observations until for no 
apparent reason the predicted deviations explode and become of the order 
of the observed values (ie the error bars get __really__wide when one 
plots pred,dev,2,*,+ or - [All this is copied from Mr Brutlags wonderful 
paper]).

eg for observations of 800 (GAUGE data), some of the DEVPREDICT values 
are
                        <!-- 2002-08-28 13:55:00 EST / 1030506900 --> 
<row><v> 2.3492396860e+02 </v></row>
                        <!-- 2002-08-28 14:00:00 EST / 1030507200 --> 
<row><v> 2.4671346109e+02 </v></row>
                        <!-- 2002-08-28 14:05:00 EST / 1030507500 --> 
<row><v> 2.5491747814e+02 </v></row>
                        <!-- 2002-08-28 14:10:00 EST / 1030507800 --> 
<row><v> 2.6110727802e+02 </v></row>
                        <!-- 2002-08-28 14:15:00 EST / 1030508100 --> 
<row><v> 2.6512894912e+02 </v></row>
                        <!-- 2002-08-28 14:20:00 EST / 1030508400 --> 
<row><v> 2.6687337106e+02 </v></row>
                        <!-- 2002-08-28 14:25:00 EST / 1030508700 --> 
<row><v> 2.6624233865e+02 </v></row>
                        <!-- 2002-08-28 14:30:00 EST / 1030509000 --> 
<row><v> 2.6315427353e+02 </v></row>
                        <!-- 2002-08-28 14:35:00 EST / 1030509300 --> 
<row><v> 2.5756715269e+02 </v></row>
                        <!-- 2002-08-28 14:40:00 EST / 1030509600 --> 
<row><v> 2.5208556860e+02 </v></row>
                        <!-- 2002-08-28 14:45:00 EST / 1030509900 --> 
<row><v> 2.4204519973e+02 </v></row>
                        <!-- 2002-08-28 14:50:00 EST / 1030510200 --> 
<row><v> 2.3070868314e+02 </v></row>

The divergence of the error bars/devpredict vals is smooth (not 
catastrophic) with a max and min on both sides (the error bar snake 
has a big lump in it - like it's ingested a squirrel).

I am using the Beta that appears not to have changed in the last 2-3 
weeks.

Should the graph or the rrd be useful, they can be sent.

Thank you,

Yours sincerely.


-- 
------------------------------------------------------------------------
Stanley Hopcroft
------------------------------------------------------------------------

'...No man is an island, entire of itself; every man is a piece of the
continent, a part of the main. If a clod be washed away by the sea,
Europe is the less, as well as if a promontory were, as well as if a
manor of thy friend's or of thine own were. Any man's death diminishes
me, because I am involved in mankind; and therefore never send to know
for whom the bell tolls; it tolls for thee...'

from Meditation 17, J Donne.

--
Unsubscribe mailto:rrd-developers-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-developers-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-developers
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi