Example 2-2 describes USER_MINING_MODEL_ATTRIBUTES
and shows a sample query. Attributes are the predictors or conditions that are used to create models and score data.
Example 2-2 USER_MINING_MODEL_ATTRIBUTES
describe user_mining_model_attributes Name Null? Type -------------------------------- -------- ---------------------- MODEL_NAME NOT NULL VARCHAR2(128) ATTRIBUTE_NAME NOT NULL VARCHAR2(128) ATTRIBUTE_TYPE VARCHAR2(11) DATA_TYPE VARCHAR2(106) DATA_LENGTH NUMBER DATA_PRECISION NUMBER DATA_SCALE NUMBER USAGE_TYPE VARCHAR2(8) TARGET VARCHAR2(3) ATTRIBUTE_SPEC VARCHAR2(4000)
The following query returns the attributes of an SVM classification model named T_SVM_CLAS_SAMPLE
in the user's schema. The model has both categorical and numerical attributes and includes one attribute that is unstructured text.
SELECT model_name, attribute_name, attribute_type, target FROM user_mining_model_attributes WHERE model_name IN 'T_SVM_CLAS_SAMPLE'; MODEL_NAME ATTRIBUTE_NAME ATTRIBUTE_TYPE TAR -------------------- ------------------------- -------------------- --- T_SVM_CLAS_SAMPLE PRINTER_SUPPLIES NUMERICAL NO T_SVM_CLAS_SAMPLE CUST_INCOME_LEVEL CATEGORICAL NO T_SVM_CLAS_SAMPLE AGE NUMERICAL NO T_SVM_CLAS_SAMPLE HOME_THEATER_PACKAGE NUMERICAL NO T_SVM_CLAS_SAMPLE CUST_GENDER CATEGORICAL NO T_SVM_CLAS_SAMPLE CUST_MARITAL_STATUS CATEGORICAL NO T_SVM_CLAS_SAMPLE BOOKKEEPING_APPLICATION NUMERICAL NO T_SVM_CLAS_SAMPLE EDUCATION CATEGORICAL NO T_SVM_CLAS_SAMPLE HOUSEHOLD_SIZE CATEGORICAL NO T_SVM_CLAS_SAMPLE FLAT_PANEL_MONITOR NUMERICAL NO T_SVM_CLAS_SAMPLE COUNTRY_NAME CATEGORICAL NO T_SVM_CLAS_SAMPLE OS_DOC_SET_KANJI NUMERICAL NO T_SVM_CLAS_SAMPLE COMMENTS TEXT NO T_SVM_CLAS_SAMPLE OCCUPATION CATEGORICAL NO T_SVM_CLAS_SAMPLE Y_BOX_GAMES NUMERICAL NO T_SVM_CLAS_SAMPLE BULK_PACK_DISKETTES NUMERICAL NO T_SVM_CLAS_SAMPLE YRS_RESIDENCE NUMERICAL NO T_SVM_CLAS_SAMPLE AFFINITY_CARD CATEGORICAL YES
See Also:
ALL_MINING_MODEL_ATTRIBUTES
in Oracle Database Reference