home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Shareware Overload
/
ShartewareOverload.cdr
/
educ
/
eztrend.zip
/
TREND.DOC
next >
Wrap
Text File
|
1980-01-01
|
10KB
|
233 lines
EASYTREND
DESCRIPTION:
EasyTrend is a program for forecasting future
projections based on historical data, and is especially
useful for managers who do not have extensive knowledge of
statistics.
OPERATING INSTRUCTIONS:
At the Main Menu, four available Options are displayed:
Forecast for a single specific future time - Use this
Option for any forecast where a specific figure is
needed; either a given level of activity at some unknown
time, or a unknown activity level at a specified future
time.
Forecast for a Time Range (trend analysis) - This Option
helps you to forecast a trend in future values based on
historical historical data.
The remaining two Options perform the same basic
function. With Easytrend you can store your data used in
calculations to a file, which can later be used for other
calulations. These two Options are for retrieving the data:
Load Data file for correlation analysis - Reload the
data for analysis under Option 1 - Forecast for a single
specific future time.
Load Data file for trend analysis - Reload the data for
analysis under Option 2 - Forecast for a Time Range
(trend analysis).
See the SAVE command Option for saving data and the
"Recalling Data" heading for use of these two Options.
All Options may be accessed by using the up/down cursor
keys to select the desired Option and pressing <RETURN>.
The command Options at the bottom of the program screens
may be accessed by using the left/right cursor keys and
pressing <RETURN>. The available Options are:
MAIN - Go to the Main Menu.
EXIT - Exit back to DOS.
PRINT - Print out the data table of results.
GRAPH - Display a graph (Option 2 only). There are two
available Options; Bar Graph and Point/Line
Graph. To print the graph, press the <Shift>
and <PrtSc> keys (see Printing EasyTrend Graphs
before doing this).
SAVE - Save the data for later retrieval and reuse. To
save, select SAVE, enter the filename (be sure to
specify the drive; eg., b:FILENAME.EXT) and press
<RETURN>. Files may be later retrieved using
Options three or four.
Forecast for a single specific future time:
After displaying a short text, the following information
will be requested:
The name of the unit value to be forecast (Sales, Costs,
Units needed, etc.)
The standard period for which you have historical data
(eg. Daily, Weekly, Yearly, etc.).
You will then be prompted to begin entering historical
data, starting with the oldest data first. A minimum of 5
periods is desirable for an accurate forecast. If you are
missing data for a particular period, simply press <RETURN>
without any entry. When data entry is complete, press "Q" to
continue to the next screen.
The next screen helps your verify your data entry. On
the left your entries will be displayed. Use the up/down
cursor keys to scroll up or down. You will be asked if the
data is correct. If yes, then you will go to the analysis
screen. If no, you will be prompted to correct the data.
To correct an entry, you will first be asked for the
number of the entry to change (first number on left of
column). You will then be asked for the new period number
and finally for the new figure to be entered. You will then
be asked again if the data is correct--answer no if there is
more data to change or yes to continue.
The first analysis screen will display the coefficient
of correlation. This is the degree of association between
the data figures entered and the time period. In every
forecast there is some variation in the data. The coefficent
of correlation indicates how much of this variation is
accounted for in the model. The higher the coefficent of
correlation, the more the variation in the data is explained
by the relationship between the data entered and time in the
model.
You will be presented with two Options:
1. Forecast a value at a future time period.
Enter the period number for which the forecast is
desired. The value will then appear along with the
standard error of estimate. This estimate is a
measure of how much the data entered is scattered
above or below the predicted values. The higher
the standard error of estimate the more concern
should be given that the data used may not reflect
the actual "population" data. In other words, the
data used in the model may be incorrect or less
accurate.
2. Future time period at which a specific value will
occur.
Enter the value for which the forecast is desired.
The period in which it should occur will then
appear along with the standard error of estimate.
This estimate is a measure of how much the data
entered is scattered above or below the predicted
values. The higher the standard error of estimate
the more concern should be given that the data used
may not reflect the actual "population" data. In
other words, the data used in the model may be
incorrect or less accurate.
Forecast for a Time Range (trend analysis):
After displaying a short text, the following information
will be requested:
The name of the unit value to be forecast (Sales, Costs,
Units needed, etc.)
The standard period for which you have historical data
(eg. Daily, Weekly, Yearly, etc.).
You will then be prompted to begin entering historical
data, starting with the oldest data first. A minimum of 5
periods is desirable for an accurate forecast. If you are
missing data for a particular period, simply press <RETURN>
without any entry. When data entry is complete, press "Q" to
continue.
You will then be asked to enter the desired number of
periods to be forecast. After entry, the next screen will
appear.
This "check screen" helps your verify your data entry.
On the left your entries will be displayed. Use the up/down
cursor keys to scroll up or down. You will be asked if the
data is correct. If yes, then you will go to the analysis
screen. If no, you will be prompted to correct the data.
You will first be asked for the number of the entry to
change (first number on left of column). You will then be
asked for the new period number and finally for the new
figure to be entered. You will then be asked again if the
data is correct--answer no if there is more data to change or
yes to continue.
The analysis screen will display the values of the data
on the left with the projected figures highlighted. The
up/down cursor keys may be used to scroll up and down through
the figures. The screen will also give the average of the
past number of figures, as well as the growth rate in
percent .
This data may be displayed in a graph if desired (see
the GRAPH command function). The available two graph types
are:
Bar graph (solid is historical, hollow is projected).
Point/Line graph.
Press <RETURN> or <Esc> to return to the analysis
screen. The graph may be printed if desired (see the next
section).
Printing EasyTrend Graphs:
You can also print the results of the trend analysis.
In order to print the graphs, you must have a printer capable
of handling the IBM graphics character set (see your printer
manual). To print a graph, you must load GRAPHICS.COM prior
to entering EasyTrend. This is done in the following manner:
Floppy Disk Users:
Place your DOS diskette in the default drive (usually A)
and at the DOS prompt type:
A> GRAPHICS press <RETURN>
The DOS prompt will reappear and GRAPHICS.COM will be
loaded.
HardDisk Users:
Simply load GRAPHICS.COM from your DOS directory.
After loading the DOS prompt will reappear.
Recalling Data:
You may save your working data for further analysis at a
later time (see SAVE command Option). This data can then be
recalled. "Load Data file for correlation analysis" (Option
3), recalls the data for forecasting for a single specific
future time. After entering the filename, you will be
brought to the "check data" screen for verification. "Load
Data file for trend analysis" (Option 4), recalls the data
for Forecast for a Time Range (trend analysis). After
entering the filename you will be prompted for the number of
periods to forecast.
Tips on Using EasyTrend:
EastTrend provides several powerful tools for making
future predictions based on historical data. However caution
must be advised. Any model is only as good as the data
entered; make sure that your data is as complete and as
consistant as possible. The less accurate the data, the less
accurate the predictions. This is why you should pay
particular attention to the the coefficent of correlation and
the standard error of estimate. Both these figures will help
you in determining if the data used is good and, if the model
the program builds is useful.
You should also be aware that since you are using
historical data to predicate future events, there will always
be some uncertainty in the actual outcome. Remember that two
predictive methods; judgement and intuition, should always be
used along with any mathematical model.