Intelligent Miner -- ResultAPI

Bivariate Statistics

DTD for Bivariate Statistics

This DTD defines the structure of bivariate statistics. Most of the definitions are reused from basic statistics, see element ModelStats

<!ELEMENT BivariateStats ( Extension*, 
                           MiningSchema, 
                           ModelStats?,
                           Partition* ) >
<!ATTLIST BivariateStats
 groupfield  %FIELD-NAME;  #IMPLIED
>

groupfield is a reference to (the name of) a MiningField. If none, it's just univariate statistics, in which case there won't be any s. Either frequencies are provided for subsets of records with common values, these subsets can be single discrete values or also intervals.If groupfield has discrete statistics, the names of the partitions might be these values, possibly preceded by the (group) field name. If groupfield has continuous statistics, the names of the partitions might be [left-margin; right-margin] for a bucket or LEFT_OUTLIERS, RIGHT_OUTLIERS, possibly preceded by the (group) field name. In both cases the names might also be MISSING or INVALID, possibly preceded by the (group) field name.

Webmaster