integer values, where
is the value speciﬁed with the NPOINT= option and
is the value speciﬁed by the MAXCOMP= option. The following table lists
for speciﬁc combinations of
. This is the actual
of components. Therefore, values larger than 8 for
duced to this upper range. It seems to be better to increase the value
MAXSTAGE= option when higher precision is requested.
speciﬁes an upper bound for the number of function calls in each optimization.
will be sufﬁcient to reach convergence. Larger values should be used if the it-
eration history indicates that the optimization process was close to a promising
solution but would have needed more than the speciﬁed number of function
calls. Smaller values should be speciﬁed when a faster but suboptimal solution
may be sufﬁcient.
speciﬁes an upper bound for the number of iterations in each optimization.
be sufﬁcient to reach convergence. Larger values should be used if the itera-
tion history indicates that the optimization process was close to a promising
solution but would have needed more than the speciﬁed number of iterations.
Smaller values should be speciﬁed when a faster but suboptimal solution may
speciﬁes an upper bound for the number of independent variables selected for
of the X’X matrix of the regression problem. The default is
Note, that the
matrix used for the stepwise regression takes
double precision values storage in RAM. For the default
speciﬁes an upper bound for the number of stages of estimation.
PROC DMNEURL: Approximation to PROC NEURAL
value is speciﬁed, the multistage estimation process is terminated
if the sum-of-squares residual in the component selection process changes
or when an upper range of 100 stages are processed.
treated differently. Large values for MAXSTAGE= may result in numerical
problems: the discretization error may be too large and the ﬁt criterion does no
longer improve and can actually become worse. In such a case the stagewise
process is terminated with the last good stage.
speciﬁes the number of values of the objective function inspected for the start
the result of the optimization especially when more than three components are
used. The default is MAXSTPT=250.
speciﬁes an upper bound for the number of eigenvectors made available for
only if there are memory problems for storing the eigenvectors when too many
variables are included in the analysis. The speciﬁed value for MAXVECT=
cannot be smaller than that for MINCOMP=.
If the speciﬁed value of
MAXVECT= is larger than the value for MAXROWS= it is reduced to the
value of MAXROWS=.
For interval targets and in a multiple stage process some memory consuming
signiﬁcantly depend on the size of the available RAM memory for those com-
putations. By default MEMSIZ=8 speciﬁes the availability of 8 mb of RAM
for such operations. Since other operations need additional memory not more
than 25 percent of the total amount of memory should be speciﬁed here. If
you are running out of memory during the DMNEURL run, you may actually
specify a smaller amout than the default 8 mb.
speciﬁes a lower bound for the number of components selected for predicting
MINCOMP= cannot be larger than that for MAXCOMP=. The MINCOMP=
speciﬁcation may permit the selection of components which otherwise would
be rejected by the STOPR2= option. PROC DMNEURL may override the
speciﬁed value when the rank of the
matrix is less than the speciﬁed
supresses the output of the status monitor indicating the progress made in the