Re: Forecasting Competition for Neural Networks & Computational Intelligence Methods


"Lionel B" <me@xxxxxxxxxxx> wrote in message news:sx5oh.54753$HV6.52056@xxxxxxxxxxxxxxxxxxxxxxx
Kent Paul Dolan wrote:
"A. M. G. Solo" <amgsolo@xxxxxxxxx> wrote

Objectives Forecast a set of 11 or 111 time series as accurately as
possible, using methods from computational intelligence and a consistent methodology.

Since there is a rather well known theorem of mathematics that given the first N entries of a series, the N+1st member is unpredictable even in concept,

I suppose it might be a merry jape to hold a time series prediction
competition and then supply the competitors with completely
unpredictable data. Looking at the website, however, suggests that this
is a serious effort and one might thus presume that the data will have
some statistical features (eg. some kind of auto-regressive structure)
that could enable prediction.

your scoring criteia must be "interesting".

It is explained on the website.

A friend?
.