Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!EU.net!nova.puug.pt!ciup2.ncc.up.pt!news.uminho.pt!caeiro!si508952
From: si508952@caeiro.ci.uminho.pt (miguel francisco a.p.rocha)
Subject: Times Series Prediction using NN
Message-ID: <1994Dec29.150808.21333@news.uminho.pt>
Sender: newsadm@news.uminho.pt (Network News Account)
Organization: Universidade do Minho, Braga, Portugal
Date: Thu, 29 Dec 1994 15:08:08 GMT
Lines: 22

29/Dec./1994

  We are developing a project of Times Series Prediction using Neural Networks,
 with the backpropagation algorithm, sigmoid function and with one hidden layer.
  The training is done by dividing the series by a constant (greater than the h
ighest value).

  PROBLEMS:
  
  -  we are getting better results using the BOX-Jenkins methodology or the Hol
t-Winters method.
  
  - we can't reach to any relation between the MSE(Mean Square Error) of the ne
ural network and the MSE in the prediction.

  - after some thousand iterations the network still converges but we get worst
 results.Is this some kind of Overfitting? When do we have to stop?

 Does anyone knows what is the better aproach in the use of Neural Networks for
 Time Series Prediction?  
-- 

