The HPNEURAL Procedure

INPUT Statement

INPUT variables </ <LEVEL=INT | LEVEL=NOM >
<STD=RANGE | STD=ZSCORE | STD=NONE >
<MISSING=MAP> >;

The INPUT statement identifies the variables in the input data set that are inputs to the neural network.

You can specify the following options:

LEVEL=INT | NOM

specifies the variables type. You can specify the following values:

INT: specifies that the variables are interval variables, which must be numeric.
NOM: specifies that the variables are nominal variables, also known as categorical variables, which can be numeric or character.

By default, LEVEL=INT.

STD=RANGE | ZSCORE | NONE

specifies how to standardize values of the interval input variables. You can specify the following values:

RANGE: standardizes the values to be in the range [–1, 1].
ZSCORE: standardizes the values by subtracting their mean and then dividing by their standard deviation.
NONE: does not standardize the values. This can be useful if the relative magnitude of input variable values are important.

By default, STD=RANGE.

MISSING=MAP

specifies that the missing value for nominal variables should be treated as a valid level (mapped to level 0). This option is not allowed for interval variables.

When training, you must include one or more INPUT statements. You need more than one INPUT statement when you have both interval and nominal input variables. The INPUT statement is not allowed when you do stand-alone scoring.

When you are training, any observation that has missing values for any variable is not used.

When you are performing stand-alone scoring, if an interval variable is missing, its mean (as observed during the training phase) is used. If a nominal variable is missing, all input neurons associated with the variable (one per class level, except for binary variables, which have a single neuron) are set to 0.

Last updated: July 02, 2020