CURVES will calculate a least-squares regression equation predicting an input factor (Y-axis) from another input factor (X-axis). The regression equations may use higher powers of the X-axis input in order to increase the predictive efficiency for non-linear relationships. The program will produce a scattergram of all data points as well as the curve representing the predicted scores. The regression equations may also be stored for later use in producing predicted scores for the input data or other samples.
Program Invocation
The specific syntax for program invocation varies between operating systems. This document diplays the code for the AIX version.
opt: Program Options (none currently supported)
ssss: The study number
TFf: Input Task Factor file
RE: Output Regression Equations file, required if the equations
produced are to be saved for later use
CURVES Control Card
CURVES: cc 01-06 The literal 'CURVES' study: cc 08-11 The study number for this run title: cc 13-72 The title for the report
X-Axis Factor Card
This card specifies the factor to use for the X-axis.
XVEC: cc 01-04 The literal 'XVEC' factor: cc 06-11 The factor name typ: cc 13-15 The factor type title: cc 17-76 An optional title for this factor
Y-Axis Factor Card
This card specifies the factor to use for the Y-axis.
YVEC: cc 01-04 The literal 'YVEC' factor: cc 06-11 The factor name typ: cc 13-15 The factor type title: cc 17-76 An optional title for this factor
PLOT Card
This card specifies the values used for plotting.
PLOT: cc 01-04 The literal 'PLOT'
mpwr: cc 06 Number of regression equations to compute. The highest
power of X will range from 1 to mpwr for each equation
respectively (max=8)
ipjt: cc 08 Projection option (1=project the plot beyond the range
of the observed X, default=project within range of X)
maxy: cc 10-13 Maximum Y-value for the plot. This value determines
the scaling for the plot. If the actual maximum is less
than 100, the default will be the most appropriate of 1,
2,5,10,25,50, or 100. If greater than 100, the actual
maximum is used as the default.
naux: cc 15-17 The number of auxiliary points
xbi: cc 19-26 X-axis bias value, added to all X input and auxiliary
point values.
ybi: cc 28-35 Y-axis bias value, added to all Y input and auxiliary
point values.
ndrp: cc 37-38 Maximum number of points to drop. The most deviant
data points will be removed, based on the value of "cor"
cor: cc 40-43 Minimum acceptable correlation between the actual and
predicted values of Y, used only if is non-zero
(default = .800). Auxiliary Point Cards
These cards are used only if the number of auxiliary points is specified on the PLOT card. These cards are normally used to introduce observations with extreme values in order to force upper and lower tails to extrapolate in the desired direction.
x-value: cc 01-10 The X-value for this point y-value: cc 11-20 The Y-value for this point
@codap curves - ssss tf100 curves ssss Good test of curves xvec tf0001/rmn yvec tf0001/rsd plot 8
Eight plots will be done, one for each power, using the values for the two input factors. A scattergram will be printed at each level.