Oracle Data Miner 11.2 Release Notes

March 2011

These release notes contain the following information:

  1. Oracle Data Miner 11.2 Functionality
  2. Before You Start
  3. Support
  4. Installation
  5. Migration
  6. Learn How Use Data Miner
  7. General Comments

Oracle Data Miner 11.2 Functionality

Oracle Data Mining, an Oracle Database Enterprise Edition Option, provides powerful data mining functionality to leverage data inside an Oracle Database. Oracle Data Miner is the graphical user interface for Oracle Data Mining. Oracle Data Miner is an extension to Oracle SQL Developer 3.0.

Oracle Data Mining enables users to build descriptive and predictive models to:

  • Predict customer behavior
  • Identify promising selling opportunities
  • Identify customer retention risks
  • Discover customer clusters, segments, and profiles
  • Detect anomalous behavior

For more information about Oracle Data Mining, see Oracle Data Mining on Oracle Technology Network.

Before You Start

To use Oracle Data Miner, you must connect to an Oracle database that satisfies these requirements:

  • The Oracle Data Mining option is installed. Oracle Data Mining is installed automatically when you install Oracle Database Enterprise Edition.
  • Oracle Text is installed. Oracle Text is installed automatically when you install Oracle Database Enterprise Edition.
  • Oracle XML DB is installed. Oracle XML DB is installed automatically when you install Oracle Database Enterprise Edition.
  • Data used by the cue card sets and the Oracle By Example tutorials requires the SH schema. If you install a starter database when you install Oracle Database Enterprise Edition, SH is automatically installed.

Data Miner has two components:

  • The Repository, which runs on an Oracle Database; one repository supports many clients.
  • The Client, which runs on any platform that SQL Developer supports (Windows, Mac, or Linux).

Support

You are supported by Oracle Support under your current Oracle Database Support license. Log Oracle Data Miner bugs and issues using My Oracle Support for the product.

You can discuss installation and use of Oracle Data Miner at Data Mining Forum.

Installation

Follow these steps to install Data Miner:

  • Step 1: Install Oracle 11g  Release 2 Database.
  • Download the software from Oracle Database Software Downloads at http://www.oracle.com/technetwork/database/enterprise-edition/downloads/index.html.
  • To use Oracle Data Miner, you must connect to an Oracle database that satisfies these requirements:
    • The Oracle Data Mining option is installed. Oracle Data Mining is installed automatically when you install Oracle Database Enterprise Edition.
    • Oracle Text is installed. Oracle Text is installed automatically when you install Oracle Database Enterprise Edition.
    • Oracle XML DB is installed. Oracle XML DB is installed automatically when you install Oracle Database Enterprise Edition.
    • Data used by the cue card sets and the Oracle By Example tutorials requires the SH schema. If you install a starter database when you install Oracle Database Enterprise Edition, SH is automatically installed.
  • Chapter 2 of Oracle Data Miner Administrator's Guide at http://download.oracle.com/docs/cd/E11882_01/datamine.112/e16807/toc.htm describes how to install Oracle Database Enterprise Edition on Microsoft Windows. For instructions for other platforms, see Installing and Upgrading in the Oracle Database Documentation Library.

  • Step 2: Download the SQL Developer 3.0 client from SQL Developer following the instructions provided on the downloads page.

  • Step 3: Install the Data Miner Repository by following the Oracle By Example "Setting Up Oracle Data Miner" at Install Data Miner Repository.

    You can install Data Miner from the Data Miner GUI, as described in the OBE, or using Installation Scripts that you download from the Downloads section of Oracle Data Miner 11g Release 2. Use of the scripts is optional.

    After you install SQL Developer 3.0, you can also view installation instructions as follows: select the menu item Help > Data Miner and then click "Getting Started" for installation information.

Migration

The script migrateodmr.sql migrates the Data Miner Repository from an earlier version (such as EA4) to the current version. All workflows created by users are preserved; at this release, only the repository software is updated.

This script is packaged with the Installation Scripts down loadable from the Downloads section of Oracle Data Miner 11g Release 2. For more information about this script, see the readme included with the scripts.

Learn How to Use Data Miner

Data Miner includes cue cards to illustrate basic usage. Select Help > Table of Contents >Data Miner Cue Cards in Oracle SQL Developer 3.0.

Data Miner also includes Oracle By Example (OBE) tutorials in the Oracle Learning Library at Oracle Data Mining 11g Release 2 OBE Series. The OBEs describe how to set up and use Oracle Data Miner.

General Comments

  • To revoke all the privileges granted to a user for data miner, run dropusergrants.sql, included in the Installation Scripts.

  • When you a build a model be cautious about turning Automatic Data Processing (ADP) off. Turning ADP off can led to failures.

  • If have SELECT privileges to other user tables/views, and the privilege is provided to you via a role, then you will not be able to create views based on that data. This will trigger failures when running the model build nodes and possibly other nodes as well. To avoid these failures, either create a table in your own user account using the Create Table node, or have the DBA or user grant SELECT privileges to you directly rather than through a role.

  • If a workflow appears to be running too long, you can use the Event Viewer to determine the status of the workflow. To open the Event Viewer, click the Event Viewer icon in the workflow tool bar. If the workflow is complete but the UI still shows that the workflow is running, simply close the workflow and reopen it.

  • A performance problem may result if you try to view the Settings tab in the model viewer for a model built using data with nested columns. This problem has been corrected in Oracle Database 11g Release 2 Patchset 11.2.0.2.