Predictive Analytics

Data mining can discover useful information buried in vast amounts of data. However, both the programming interfaces and the data mining expertise required to obtain these results may be too complex for use by the wide audiences that can obtain benefits from Oracle Data Mining.

The Predictive Analytics procedures Predict and Explain address both of these complexities by automating the later stages of data mining process from certain types of data preprocessing through model building to scoring new data. Predict and Explain provide an important tool that makes data mining possible for a broad audience of users, in particular, business analysts.

Data used by Oracle Data Mining consists of tables or views stored in an Oracle database. Each column in a record (row) holds an item of information. Data mining models are often used to identify important columns or to Predict column values. Explain ranks attributes in order of importance in explaining a target column. Predict predicts the value of a target.

Input for Predict or Explain is a table or view in an Oracle database. Each column in the table must have one of the following datatypes: NUMBER, FLOAT, CHAR, VARCHAR2, DATE, or TIMESTAMP.

For detailed information about Predict and Explain, see the description of DBMS_PREDICTIVE_ANALYTICS in the PL/SQL Packages and Types Reference.

Note: The Oracle Data Mining 11gPL/SQL package DBMS_PREDICTIVE_ANALYTICS supports Explain. Oracle Data Miner 11.1 does not support Explain.