ModelDetail
Objects
This interface encapsulates the representation detail of a data mining model.
m.getAlgorithmName() : string
Returns the name of the model builder algorithm.
m.getApplicationName() : string
Returns the name of the model builder application.
m.getApplicationVersion() : string
Returns the version of the model builder application.
m.getBuildDate() : Date
Returns the date this model was built.
m.getCopyright() : string
Returns the model copyright.
m.getInputDataModel() : DataModel
Returns the input data model required by the model. Note that the ModelingRole
for
each field in the input data model is ModelingRole.IN
,
even for fields that had been specified as ModelingRole.BOTH
for
association or sequence models.
m.getModelID() : string
Returns the model ID. For models within a composite model, it is assumed that each model has a unique ID.
m.getModelType() : ModelType
Returns the type of the model.
m.getOutputDataModel() : DataModel
Returns the output data model produced by the model. Note that
the ModelingRole
for each field in the output data
model is ModelingRole.OUT
, even for fields that had
been specified as ModelingRole.BOTH
for association
or sequence models.
It is also important to note that this is not same as the output
column set produced by the ModelApplier
that applies
the model to data. The transformer output column set will usually
include additional fields from the input and may include property
settings that modify the number or type of outputs produced (for example,
a transformer that applies a clustering model may produce the cluster
ID as an integer or a string depending on the property settings).
m.getPMMLModelType() : PMMLModelType
For a model which is represented internally using PMML (one for which isPMMLModel() returns
True
) this returns the PMML model type from the underlying PMML that was generated
by the modeling algorithm. For a model which is not represented using PMML this returns
None
.
m.getPMMLText() : string
For a model which is represented internally using PMML (one for which isPMMLModel() returns
True
) this returns a string representing the PMML generated from the model building
algorithm. For non-PMML models this returns None
. When this returns
None
you may be able to obtain an alternative representation of the model as PMML
using the export function: TaskRunner.exportModelToFile()
.
m.getSplitColumnNames() : List
Returns the list of split column names if this is a split model or None
for
non-split models.
m.getSplitModelCount() : int
Returns the number of split models if this is a split model or 0
otherwise.
m.getSplitModelKey(index) : List
index (int) :
the model index
Returns the values of the split fields at the specified index for split models or
None
for non-split models. When present, the values are in the order specified by
the split output columns.
m.getSplitModelPMMLText(index) : string
index (int) :
the model index
Returns the PMML representation of the split model at the specified index for split models or
None
for non-split models or if the split model is not represented as PMML.
m.getUserName() : string
Returns the name of the account used to build the model.
m.isOutputColumnAuxiliary(name) : boolean
name (string) :
the column name
Returns True
if the supplied column name is an auxiliary column in the output
data model or False
. An auxiliary output column is a column that provides
additional information about the output of the model, for example, the confidence of the prediction
or distance from the cluster center.
m.isPMMLModel() : boolean
Returns True
if this model is represented using PMML, or False
otherwise.
m.isSplitModel() : boolean
Returns True
if this model is a split model, False
otherwise.