This preview shows page 1. Sign up to view the full content.
Unformatted text preview: (hopefully) Objective Forecasting Methods
Two primary methods: causal models and time series
causal
and
methods
methods
Causal Models:
Let Y be the quantity to be forecasted and (X1,
X2, . . . , Xn) be n variables that have predictive power
be
for Y. A causal model is
Y = f (X1, X2, . . . , Xn).
(X
).
A typical relationship is a linear one. That is,
typical
Y = a0 + a1X1 + . . . + an Xn. 4 Time Series Methods
A time series is just collection of past values of the
time
variable being predicted. Also known as naive methods.
Goal is to isolate patterns in past data.
Trend
Trend
Seasonality
Seasonality
Cycles
Cycles
Randomness
Randomness Time Series Patterns 5 Notation Conventions
Let D1, D2, . . . Dn, . . . be the past values of the series to
Let
be
be predicted (observed demand). If we are making a
forecast in period t, assume we have observed Dt , Dt1
etc.
etc.
Let Ftt, t + τ be forecast made in period t for the demand in
Let ,
forecast
period t + τ where τ = 1, 2, 3, …
1,
Then Ft 1, t is the forecast made in t1 for t and
Ft, t+1
Then 1, is
iis the forecast made in t for t+1. (one step ahead) Use
s
shorthand notation Ft = Ft  1, t .
shorthand Evaluation of Forecasts
Evaluation
The forecast error in period t, et, iis the difference
s
between the forecast for demand in period t and the
actual value of demand in t.
For a multiple step ahead forecast:
et = Ft  τ, t Dt.
For one step ahead forecast: et = Ft  Dt.
For
MAD = (1/n) Σ  ei 
MSE = (1/n) Σ ei 2 6 Example
Compare the accuracy of the forecasts of two managers
Compare
of a SRAM manufac...
View
Full
Document
This document was uploaded on 03/23/2014.
 Spring '14

Click to edit the document details