4.9 Model Settings in the Data Dictionary

Explains about ALL/USER/DBA_MINING_MODEL_SETTINGS in data dictionary view.

Information about Oracle Machine Learning model settings can be obtained from the data dictionary view ALL/USER/DBA_MINING_MODEL_SETTINGS. When used with the ALL prefix, this view returns information about the settings for the models accessible to the current user. When used with the USER prefix, it returns information about the settings for the models in the user's schema. The DBA prefix is only available for DBAs.

The columns of ALL_MINING_MODEL_SETTINGS are described as follows and explained in the following table.

SQL> describe all_mining_model_settings
 Name                                      Null?    Type
 ----------------------------------------- -------- ----------------------------
 OWNER                                     NOT NULL VARCHAR2(30)
 MODEL_NAME                                NOT NULL VARCHAR2(30)
 SETTING_NAME                              NOT NULL VARCHAR2(30)
 SETTING_VALUE                                      VARCHAR2(4000)
 SETTING_TYPE                                       VARCHAR2(7)

Table 4-84 ALL_MINING_MODEL_SETTINGS

Column	Description
owner	Owner of the machine learning model.
model_name	Name of the machine learning model.
setting_name	Name of the setting.
setting_value	Value of the setting.
setting_type	INPUT if the value is specified by a user. DEFAULT if the value is system-generated.

The following query lists the settings for the Support Vector Machine (SVM) classification model SVMC_SH_CLAS_SAMPLE. The ALGO_NAME, CLAS_WEIGHTS_TABLE_NAME, and SVMS_KERNEL_FUNCTION settings are user-specified. These settings have been specified in a settings table for the model. The SVMC_SH_CLAS_SAMPLE model is created by the oml4sql-classification-svm.sql example.

Example 4-16 ALL_MINING_MODEL_SETTINGS

SQL> COLUMN setting_value FORMAT A35
SQL> SELECT setting_name, setting_value, setting_type
            FROM all_mining_model_settings
            WHERE model_name in 'SVMC_SH_CLAS_SAMPLE';
 
SETTING_NAME                   SETTING_VALUE                       SETTING
------------------------------ ----------------------------------- -------
SVMS_ACTIVE_LEARNING           SVMS_AL_ENABLE                      DEFAULT
PREP_AUTO                      OFF                                 DEFAULT
SVMS_COMPLEXITY_FACTOR         0.244212                            DEFAULT
SVMS_KERNEL_FUNCTION           SVMS_LINEAR                         INPUT
CLAS_WEIGHTS_TABLE_NAME        svmc_sh_sample_class_wt             INPUT
SVMS_CONV_TOLERANCE            .001                                DEFAULT
ALGO_NAME                      ALGO_SUPPORT_VECTOR_MACHINES        INPUT