• To: mathgroup at smc.vnet.net
• Subject: [mg7672] HELP with exponential smoothing, PLEASE!
• From: arosa at mail.telepac.pt (Antonio Rosa)
• Date: Thu, 26 Jun 1997 01:36:56 -0400 (EDT)
• Organization: PORTUGAL TELECOM
• Sender: owner-wri-mathgroup at wolfram.com

```        Hi !, i'm using an programm called STATISTICA to forecast
telephones demand per month. I've acheived good results using an
linear trend in exponential smoothing, but now i'm trying to develop
an program to make the same calculations.....but i'm getting different
results. Can you help me? here are the results from the program
STATISTICA:

month: JANUARY
S0 (estimated from data) = 341884
T0 (estimated from data) = 102675
alpha=0.55 ; beta=0.1  (constants)
case   Real Values    Smoothed Series    Resids
1        393221            444559                   -51337.5
2        562774            614586                    2380.3
3        611754            682960                   -71206.1
4        803921            736751                   67169.7
5                                870344 (* forecasted)

I'm using the following formulas:

T0=(Xn-X1)/(N-1) , where N is the length of series
S0=X1-(T0/2)

Formula of trend:   T(t)=beta* ( F(t) - F(t-1)  ) + (1-beta)*T(t-1)  ,
where T is trend and  F is forecast

F(t+1)=alpha*( A(t) ) + (1-alpha)*( F(t-1) - T(t-1) )  , where A is
the real value

I think that this last formula is wrong, can anyone tell me the
correct formula(s) so that i can get the same values of STATISTICA ?

Antonio Rosa

----
Antonio Rosa
Portugal Telecom
Special Systems and Services
Coimbra - Portugal
mail:  arosa at mail.telepac.pt

```

• Prev by Date: nested * and ** (rules for commutative quantities)
• Next by Date: using indexed variables in MultiplierMethod[ ], etc.
• Previous by thread: nested * and ** (rules for commutative quantities)
• Next by thread: using indexed variables in MultiplierMethod[ ], etc.