Re: the best forecasting method

From: Phil Sherrod (phil.sherrod_at_REMOVETHISsandh.com)
Date: 06/16/04 

Date: Wed, 16 Jun 2004 14:32:33 GMT

On 15-Jun-2004, carrotcheung@hotmail.com (C. Cheung) wrote:

> I would like to select a forecasting method for a set of data, here
> are my list of forecasting methods:
> 1. Moving Average 3
> 2. Winters' Exponential Smoothing
> 3. ARIMA
> 4. SARIMA
> 5. Neural Network
> Other than evaluating the method by MAPE, Theil's U....etc.
> Is there any other hints I can get before trying every method, for
> example, from the pattern or trend of the data???
> Is each of the method above perform best with a particular data
> pattern?

The best method will depend on the type of data you are trying to model. If
you are modeling a scientific/engineering application such as the measured
position of a pendulum versus time, then nonlinear regression would be
appropriate. But if you are trying to predict probable purchases by
customer categories based on demographic information, then decision trees or
possibly neural networks would be more appropriate.

So please describe in more detail the nature of the data that you are trying
to model. 

-- 
Phil Sherrod
(phil.sherrod 'at' sandh.com)
http://www.dtreg.com  (decision tree modeling)
http://www.nlreg.com  (nonlinear regression)

Relevant Pages

  • Re: Question about Decision Trees and Neural Networks
    ... Avg and Amount is long Continous attributes. ... of all attributes to continuous int/double for the Neural Network. ... I get many predictions that are negative values even though all ... and for Decision Trees with the same data I get around 110. ...
    (microsoft.public.sqlserver.datamining)
  • Question about Decision Trees and Neural Networks
    ... And another with the same 9 attributes + a avarage Amount for all combinations of the nine attribute for every transaction. ... I have alot of transactions and a read somewhere that a Neural Network should work better than a decision tree in a case similar to mine. ... The score in the "Lift chart" for the Neural Network model becomes 0,00 and for Decision Trees with the same data I get around 110. ...
    (microsoft.public.sqlserver.datamining)
  • Gini coefficient
    ... neural network without using enterprise miner (although I have ... enterprise miner licence). ...
    (sci.stat.consult)