supresses all output printed in the output window.
number of discretization points (should be even inbetween 5 and 19). By de-
in the model using the MINCOMP= and MAXCOMP= options.
speciﬁes the criterion for the optimization:
target the Goodman-Kruskal
is applied on a frequency table deﬁned by
criterion is minimzed.
When this option is speciﬁed the weight must be speciﬁed using the
CRITWGT= option. Currently this option is valid only for binary target.
If an OUTSTAT= data set is speciﬁed, i.e. only principal components are
If PMATRIX is speciﬁed, the
, the covariance, or the correlation
If no OUTSTAT= data set is speciﬁed, i.e. a nonlinear model based on
trates the output options:
This option is valid only if an OUTSTAT= data set is speciﬁed, i.e. when
DMNEURL is used only for computing eigenvalues and eigenvectors of the
, covariance, or correlation matrix. If PMATRIX is speciﬁed, this matrix
by that of the PALL option.
print the detailed histories of all optimization processes. The PALL option
includes only the summarized forms of the history output (header and result).
see the PALL option for the amount of output being printed.
speciﬁes the output of accuracy tables. This option is invoked automatically if
the PALL option is speciﬁed.
PROC DMNEURL: Approximation to PROC NEURAL
speciﬁes the criterion for selecting the best result among all of the activation
SELCRIT=SSE select solution with smallest sum-of-squares error.
SELCRIT=ACC select solution with largest accuracy rate. (For interval tar-
get the Goodman-Kruskal
ror. This option is valid only for binary target. When this option is speci-
ﬁed the weight must be speciﬁed using the CRITWGT= option.
speciﬁes a criterion for the singularity test. The default is
should not be changed if there are no signiﬁcant reasons to do so.
speciﬁes a lower value for the incremental model
value at which the variable
values of the components selected in the range speciﬁed by the MINCOMP=
and MAXCOMP= values. The default is
permits the use of a data set generated by PROC DMDB to be speciﬁed as a
TESTDATA= input data set. If this option is not speciﬁed, the data set speciﬁed
with TESTDATA= must be a normal SAS data set.
For the syntax of the DECISION statement see the document of PROC DECIDE.
An activation function
and a link function
may be speciﬁed for the mapping inbe-
tween the component scores
and the values
of the response variable (stage=0)
(or the residuals in stage > 0),
for each activation function
. The FUNCTION and LINK state-
ment can be used to specify the functions
If more than one function
is speciﬁed, each of the speciﬁed functions is
squares residual or accuracy (see SELCRIT= option) is selected. By default all
available activation functions are used.
LINK statement Currently only one of the following link functions can be used for
the outer function :
function is used for interval target. In a parallelized version of PROC DM-
NEURL, multiple functions
could be feasible.
One variable name may be speciﬁed identifying the target (response) variable for the
two regressions. Note, that one or more target variables may be speciﬁed already
with the PROC DMDB run. If a target is speciﬁed in the PROC DMDB run, it must
not be speciﬁed in the PROC DMNEURL call.
VAR or VARIABLES Statement
VAR varlist ;
VARIABLES varlist ;
All variables, numeric (interval) and categorical (CLASS) variables which may be
used for independent variables are speciﬁed with the VAR statement.
FREQ or FREQUENCY Statement
FREQ onevar ;
FREQUENCY onevar ;
One numeric (interval scaled) variable may be speciﬁed as a FREQ variable. Note,
that a rational value is truncated to the next integer. It is recommended to specify
the FREQ variable already in the PROC DMDB run. Then the information is saved
in the catalog and that variable is used automatically as a FREQ variable in PROC
DMNEURL. This also ensures that the FREQ variable is being used automatically
by all other PROCs in the EM project.