Changes in this release for Oracle Data Mining User’s Guide.
This guide is new in release 12c. Oracle Data Mining User's Guide replaces two manuals that were provided in previous releases: Oracle Data Mining Administrator's Guide and Oracle Data Mining Application Developer's Guide.
Information about database administration for Oracle Data Mining is now consolidated in Administrative Tasks for Oracle Data Mining . The remaining chapters of this guide are devoted to application development.
Information about the Data Mining sample programs is now in The Data Mining Sample Programs.
The following changes are documented in Oracle Data Mining User’s Guide for 12c Release 2 (12.2).
The following features are new in this release:
Partitioned Models
Data Mining SQL function
A new Data Mining SQL function ORA_DM_PARTITION_NAME is included for partitioned models. The function returns the partition names for a partitioned model.
See Data Mining SQL Scoring Functions.
Provided new scoring functions
See Partitioned Model scoring.
See GROUPING Hint.
About partitioned model
Description of Partitioned model is added
DDL in partitioned model
Explained the newly added Add and Drop partition for maintenance operations.
Model Views
Added new Model Detail Views. Model Detail Views are preferred over GET* functions.
See Model Detail Views.
New Data Dictionary Views. See Data Mining Data Dictionary Views.
Explicit Semantic Analysis
Newly added FEATURE_COMPARE SQL function
FEATURE_COMPARE SQL function
Provides an example of the new SQL function FEATURE_COMPARE using ESA algorithm.
Association Rules Aggregates
Using retail analysis data
Added enhancements to Association Rules and an example to show the concept of aggregates.
See Using Retail Analysis Data.
See Model Detail Views.
R Extensibility
Mining model settings for R
New mining model settings are included for R, to define the characteristics of R models. The mining model settings can be used with generic settings that are independent of algorithms, to specify R model build, score and view.
DBMS_DATA_MINING for R
The DBMS_DATA_MINING subprograms that are independent of algorithms, can operate on R model for mining functions such as Classification, Clustering, Feature Extraction, and Regression.
See DBMS_DATA_MINING.
The following changes are documented in Oracle Data Mining User's Guide for 12c Release 1 (12.1).
The following features are new in this release:
Expanded prediction details
The PREDICTION_DETAILS function now supports all predictive algorithms and returns more details about the predictors. New functions, CLUSTER_DETAILS and FEATURE_DETAILS, are introduced.
See Prediction Details.
Dynamic scoring
The Data Mining SQL functions now support an analytic clause for scoring data dynamically without a pre-defined model.
See Dynamic Scoring.
Significant enhancements in text mining
This enhancement greatly simplifies the data mining process (model build, deployment and scoring) when unstructured text data is present in the input.
Manual pre-processing of text data is no longer needed.
No text index must be created.
Additional data types are supported: CLOB, BLOB, BFILE.
Character data can be specified as either categorical values or text.
New clustering algorithm: Expectation Maximization
See the following:
New feature extraction algorithm: Singular Value Decomposition with Principal Component Analysis
See the following:
Generalized Linear Models are enhanced to support feature selection and creation.
The following features are no longer supported by Oracle. See Oracle Database Upgrade Guide for a complete list of desupported features in this release.
Oracle Data Mining Java API
Adaptive Bayes Network (ABN) algorithm
The following are additional new features in this release:
A new SQL function, CLUSTER_DISTANCE, is introduced. CLUSTER_DISTANCE returns the raw distance between each row and the cluster centroid.
See Scoring and Deployment .
New support for native double data types, BINARY_DOUBLE and BINARY_FLOAT, improves the performance of the SQL scoring functions.
See Preparing the Data.
Decision Tree algorithm now supports nested data.
See Preparing the Data.