4.2.1.3 Specify Model Settings

Functions in the OREdm package have an argument that specifies settings for an Oracle Machine Learning for SQL model and some have an argument for setting text processing parameters.

General Parameter Settings

With the odm.settings argument to an OREdm function, you can specify a list of OML4SQL parameter settings. Each list element's name and value refer to the parameter setting name and value, respectively. The setting value must be numeric or string. Refer to Specify Model Settings in Oracle Machine Learning for SQL User’s Guide for each algorithm's valid settings.

The settings function returns a data.frame that lists each OML4SQL parameter setting name and value pair used to build the model.

Text Processing Attribute Settings

Some OREdm functions have a ctx.settings argument that specifies text processing attribute settings with which you can specify Oracle Text attribute-specific settings. With the odm.settings argument, you can specify the Oracle text policy, the minimal number of documents in which each token occurs, and the maximum number of distinct features for text processing. With the ctx.settings argument, you specify the columns that should be treated as text and the type of text transformation to apply.

The ctx.settings argument applies to the following functions:

• ore.odmESA, Explicit Semantic Analysis

• ore.odmGLM, Generalized Linear Models

• ore.odmKMeans, k-Means

• ore.odmNMF, Non-Negative Matrix Factorization

• ore.odmSVD, Singular Value Decomposition

• ore.odmSVM, Support Vector Machine

Note:

To create an Oracle Text policy, the user must have the CTXSYS.CTX_DDL privilege.

See Also:

Create a Model that Includes Machine Learning Operations on Text in Oracle Machine Learning for SQL User’s Guide for valid text attribute values.