data mining capabilities through OJDM or the SQL-based interfaces.
SQL operators in particular can enable complex data mining deploy-
ment scenarios, such as scoring multiple models in a single query or
immediately combining the results of mining with other database
SQL features involving spatial data and text analysis. For more
details about these APIs, refer to [ORADMAPPDOC 2006].
Oracle Data Mining PL/SQL packages and SQL functions
Provides the PL/SQL procedures and functions that can be used to build,
apply, and test mining models.
Provides the PL/SQL transformation procedures for binning, clipping,
normalization, and missing value handling.
Provides the predict and explain procedures that automate complex data
mining processes and produces final results.
These prediction operators are used to compute prediction-related values
using classification and regression models built using either OJDM or
DBMS_DATA_MINING. The prediction operator returns the top
prediction value of a case. The prediction_probablity operator returns
the probability value for the top prediction. The prediction_cost
operator returns the cost value of the top prediction if the cost matrix is
specified at the model build. The prediction_set operator is used with
classification models to return the collection of prediction, probability
and cost values for all the class values of the target attribute in the model
build data. The prediction_details operator returns the XML
representation of the model details that describes how the prediction has
been made for each case.
These cluster operators are used to assign clusters for each case using
the clustering models build using OJDM or DBMS_DATA_MINING.
The cluster_id operator returns the case's top cluster id and
cluster_probability operator returns the probability associated with
the top cluster. The cluster_set operator returns the collection of cluster
ids and their associated probabilities for all clusters in the model.
These feature operators are used to find the feature associated with the
case using the feature extraction model built using OJDM or
DBMS_DATA_MINING. The feature_id operator returns the case's
feature id that has highest coefficient value. The feature_value operator
returns the coefficient value associated with the top feature for the case.
The feature_set operator returns the collection of feature id and
associated coefficient value for all features in the model.