RULEGENERATION subcommand (SPATIAL ASSOCIATION RULES command)
The RULEGENERATION
subcommand specifies parameters for how rules are
generated. A rule consists of conditions and predictions. A rule can have multiple conditions
and multiple predictions.
- MAXCONDITION=integer
- Maximum number of field values that can occur in the conditions for a rule. The value
must be a positive integer between 1 and 9. The default value is 5. The sum of
MAXCONDITION
andMAXPREDICTION
cannot exceed 10. - MAXPREDICTION=integer
- Maximum number of field values that can occur in the predictions for a rule. The value
must be a positive integer between 1 and 5. The default value is 1. The sum of
MAXCONDITION
andMAXPREDICTION
cannot exceed 10. - MINSUPPORT=value
- Minimum rule support a rule must have to be included in the output. The value represents the proportion of cases for which the rule is true in the observed data. For example, if the rule is "If x=1 and y=2, then z=3," rule support is the actual proportion of cases in the data for which x=1, y=2, and z=3. The value must be greater than 0 and less than 1. The default value is 0.05. Using the default value, a rule must be true at least 5% of the time in the observed data.
- MINCONDITIONSUPPORT=value
- Minimum condition support a rule must have to be included in the output. The value represents the proportion of cases for which the condition exists. For example, if the rule is "If x=1 and y=2, then z=3," condition support is the proportion of cases in the data for which x=1 and y=2. The value must be greater than 0 and less than 1. The default value is 0.05. Using the default value, a condition must occur in 5% of the cases.
- MINCONFIDENCE=value
- Minimum confidence a rule must have to be included in the output. Confidence is the proportion of correct predictions. The value must be greater than 0 and less than 1. The default value is 0.1. Using the default value, the prediction of the rule must be correct at least 10% of the time.
- MINLIFT=value
- Minimum lift a rule must have to be included in the output. Lift is a measure of how much the rule improves prediction compared to random chance. It is the ratio of correct predictions to the overall occurrence of the predicted value. The value must be greater than 1. The default value is 1.2. For example, if the predicted value occurs 20% of the time and the confidence in the prediction is 80%, then the lift value is 4. The default value of 1.2 means that prediction must be correct 20% more often than the overall occurrence of the predicted value.
- EXCLUDE=fieldname(id) fieldname(id); fieldname(id) fieldname(id);...
- Excludes the specified pairs of fields from being included in the same rule.
- Each field name is followed by the ID value of the data source it comes from.
- The ID value is the
IDENTIFIER
value that is specified on theDATA
subcommand of theSPATIAL MAPSPEC
command. - If there are multiple event data sources, use the keyword
EVENT
for the ID value. - Each pair of field names is separated by a semi-colon.
- ENCODEASSAME=fieldname(id) fieldname(id); fieldname(id) fieldname(id);...
- Identifies pairs of fields that should be treated as the same field.
- Each field name is followed by the ID value of the data source it comes from.
- The ID value is the
IDENTIFIER
value that is specified on theDATA
subcommand of theSPATIAL MAPSPEC
command. - If there are multiple event data sources, use the keyword
EVENT
for the ID value. - Each pair of field names is separated by a semi-colon.