home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Between Heaven & Hell 2
/
BetweenHeavenHell.cdr
/
500
/
484
/
s.gth
< prev
next >
Wrap
Text File
|
1986-09-07
|
7KB
|
115 lines
[S] [STATISTICS] Command
Subcommands:
Arith. Const. Inflation Lead/lag Moving-ave. Period Rate Smooth Trend
Introduction:
The statistics subcommands always operate on data held in memory 0 then write
the final result back to memory 0. If the results are not required then
immediately press UNDO (Fn key 4) to restore memory 0 and the screen back to
its previous data values. Some of the subcommands automatically produce a new
line graph of the results (e.g. Trend and Smooth). Most of the others leave
results in memory 0 for subsequent graphing.
Arithmetic [SAx] (x = File or Sum)
File-calcs: Enables calculation between memory 0 and another memory or
disk file. All matching X values are used in the calculations.
The operands available are: + - * /
e.g.s. SAF+SALES<RTN> SAF*NEWRATE<RTN> SAF-DISCOUNT<RTN>
Sum-mem-0: Each succeeding data value is added to produce a cumulative
data file in memory 0, starting at a specified date.
e.g. SAS10/86<RTN> (Sum data from Oct. 1986)
Constant [SCxn] (x = Fixed or Season. n = value or date)
Fixed-constant: A constant may be used to modify the data in memory 0
with the operators: + - * / and start at a specified date.
(The default date is the start date.)
e.g. SCF+8<RTN> 5/86<RTN> (Add 8 to all data values
from May 1986.)
Season-adjust: All matching X values in memory 0 will be multiplied by a
set of seasonal factors held in a disk file. This season
file is created with the normal file create [FCD] command.
The filename must start with "SEA" followed by a number in
the range 0 - 9. e.g. SEA4
(The Season file usually contains 12 monthly values.)
The first date in the Season file must match the first date
in the data file.
e.g. SCS2 This takes the factors in file "SEA2"
and multiplies matching data values currently in mem 0.
Inflation [SInd] (n = % inflation. d = starting date)
This command modifies the data in memory 0 with a specified inflation rate
from a specified date. SI command is only valid with time series data.
Eg. SI11<RTN> 7/86<RTN> Calculates the effect of 11% inflation from
July 1986.
If a date is not entered (press <RTN>) then the first date is the default.
Leadlag [SLxn] (x = leaD or laG n = number of time periods)
leaD [SLDn]: The starting point or date is moved forward to an earlier
time by a specified number of time periods on the X axis.
The data itself is not altered.
Eg. SLD4<RTN> (Move starting date forward 4 time periods)
laG [SLGn]: The starting point or date is moved backward to a later
time by a specified number of time periods on the X axis.
The data itself is not altered.
Eg. SLG3<RTN> (Move starting date later 3 time periods.)
Moving-average [SMn] (n = number of time periods)
This command takes the average of the specified (n) number of time
periods data. The result is placed back in memory 0 and the moving-ave
line is graphed.
Period [SPx] (x = Month, Quarter, Year or Other)
This forces time-series data in memory 0 to be converted to Monthly,
Quarterly or Yearly data.
For example, if the original file contained monthly data
and the command SPQ is given, each calendar three months will be summed
to give quarterly data.
In converting quarterly data to monthly, each data value is divided
by three to give "average" monthly figures.
Egs. SPM SPQ SPY
If "Other" is selected then any data can be grouped by any number from
2 up to the maximum number of points. In addition, the data points in
each group can either be totalled or averaged.
e.g. SPO4<RTN>T will total each group of 4 data points to produce a new
data point in memory 0.
e.g. SPO3<RTN>A will average each group of 3 data points to produce a new
data point in memory 0.
Rate [SRx] (x = Actual or Percentage)
This computes the rate of change of the data from each X-axis period to
the next. (The data is differentiated.) The results are placed in memory 0
As the scale on the Y-axis is likely to be altered, the scales are auto-
matically recalculated and the rate-of-change line is graphed after the
screen is cleared. e.g.s SRP or SRA then G<RTN>
Smooth [SSxn] (x = Normal, Double or Triple. n = smoothing factor)
Data values varying rapidly over a short range can often obscure the
long term direction. The smooth command produces a curve following the
general direction of the values. The smoothing factor can be in the
range: 0.01 (extreme smoothing) to .99 (almost no smoothing). By using
the UNDO key, you can experiment with different smoothing factors until
the required result is obtained. Note that the smoothed data values
are written to memory 0. Double and Triple smoothing options produce
slightly different smooth curves which may better fit different patterns
of unsmoothed data.
e.g. SSN .3<RTN> (Normal smoothing with a .3 factor.)
(A smoothing factor >1 will produce a graph with larger deviation
than the original. This could be used to highlight small changes.)
Trend [STxn] (x = Normal, Logarithmic or Exponential. n = starting date)
Normal Trend [STNn]: A straight line is derived from the data points
and graphed as a linear least-squares trend.
Logarithmic Trend [STLn]: Computes a logarithmic curve-fit trend line. A
log curve rises steeply at first with gradually decreasing slope.
Exponential Trend [STEn]: Computes an exponential-fit trend line. (Avoid
zero or negative values.)
The starting date is optional and can be used to show short-term trends.
To produce more accurate trend curve-fitting, the original data can be
modified with a constant before calculating the trend.
A trend line could be further modified with a Seasonal-factor file to
show likely actual values. (e.g. projected sales.)
e.g. STN 7/87<RTN> (Calculate straight-line trend from July 1987.)
September 23, 1986