Professional Documents
Culture Documents
Outline
Personnel
Workforce planning,
hiring, layoff
Characteristics of Forecasts
Forecasts are always
wrong; so consider both
expected value and a
measure of forecast error.
Long-term forecasts are
less accurate than shortterm forecasts. For
example, 7-Eleven Japan
Characteristics of Forecasts
Some time series (called aggregate series) are obtained
by summing up more than one time series (called
disaggregate series) . For example, annual sales are
obtained by adding 12 monthly sales. The annual sales
is an aggregate series and monthly sales is a
disaggregate series. Aggregate forecasts are more
accurate than disaggregate forecasts. Variation in GDP
of a country is much less than the annual earnings of a
company. Consequently, it is easy to forecast the GDP
of a country with less than 2% error. However, it is much
more difficult to forecast annual earning of a company
with less than a 2% error.
Components of Demand
Average demand
Trend
Gradual shift in average demand
Seasonal pattern
Periodic oscillation in demand which repeats
Cycle
Similar to seasonal patterns, length and magnitude
of the cycle may vary
Random movements
Auto-correlation
Qantity
Components of Demand
Time
(a) Average: Data cluster about a horizontal line.
Quantity
Components of Demand
Time
(b) Linear trend: Data consistently increase or decrease.
Components of Demand
Quantity
Year 1
Months
(c) Seasonal influence: Data consistently show
peaks and valleys.
Components of Demand
Quantity
Year 1
Year 2
|
Months
(c) Seasonal influence: Data consistently show
peaks and valleys.
Components of Demand
Quantity
Components of Demand
Years
(c) Cyclical movements: Gradual changes over
extended periods of time.
Components of Demand
Suppose that a company
has institute a sales
incentive system that
provides a bonus for the
employee with the best
improvement in bookings
from one month to the next.
With such an incentive in
place, a month of poor sales
is often followed by a month
of good sales. Similarly, a
month of good sales would
usually be followed by a lull.
Components of Demand
This means that sales in
consecutive months tend to
be negatively correlated.
This information can be
used to improve sales
forecasts. Autocorrelation is
the correlation among
values of observed data
separated by a fixed number
of periods. In the example
above, we would say that
the series has a negative
autocorrelation of order one.
Trend
Random
movement
Time
Demand
Demand
Components of Demand
Trend with
seasonal pattern
Time
Snow Skiing
Seasonal
Long term growth trend
Evaluation of Forecast
There are many forecasting techniques and software. The
performance of a forecasting technique can be measured by
the error produced over time.
We shall now discuss various measures used to evaluate
forecasting techniques. Let,
Dt Actual data in period t
F t Forecast in period t
Et Forecast error in period t Ft Dt
Standard deviation of forecast errors
MSE Mean Squared Error
MAD Mean Absolute Deviation
MAPE Mean Absolute Percentage Error
Evaluation of Forecast
MSE =
=
2
t
n
MSE
MAD =
MAPE =
|Et |
n
[ |Et | (100) ] / Dt
n
Evaluation of Forecast
Month,
t
Demand,
Dt
Forecast,
Ft
1
2
3
4
5
6
7
8
200
240
300
270
230
260
210
275
225
220
285
290
250
240
250
240
Total
Error,
Et
Absolute
Error
Absolute
Percent
Squared, Error,
Error,
2
Et
|Et|
(|Et|/Dt)(100)
Evaluation of Forecast
Measures of Error
MSE =
MAD =
MAPE =
=
=