Control Access to Oracle Machine Learning for SQL Models and Data |
您所在的位置:网站首页 › access privileges › Control Access to Oracle Machine Learning for SQL Models and Data |
The CREATE MINING MODEL is a privilege that you must have to create and perform operations on your model. Some other machine learning privileges can be assigned by issuing GRANT statements. You must have the CREATE MINING MODEL privilege to create models in your own schema. You can perform any operation on models that you own. This includes applying the model, adding a cost matrix, renaming the model, and dropping the model. The GRANT statements in the following example assign a set of basic machine learning privileges to the oml_user account. Some of these privileges are not required for all machine learning activities, however it is prudent to grant them all as a group. Additional system and object privileges are required for enabling or restricting specific machine learning activities. The following table lists the system privileges required for running the OML4SQL examples. Table 7-2 System Privileges Granted by dmshgrants.sql to the OML4SQL User Privilege Allows the OML4SQL User ToCREATE SESSION Log in to a database session CREATE TABLE Create tables, such as the settings tables for CREATE_MODEL CREATE VIEW Create views, such as the views of tables in the SH schema CREATE MINING MODEL Create OML4SQL models EXECUTE ON ctxsys.ctx_ddl Run procedures in the ctxsys.ctx_ddl PL/SQL package; required for text mining Example 7-7 Privileges Required for Machine Learning This example grants the required privileges to the user oml_user. GRANT CREATE SESSION TO oml_user; GRANT CREATE TABLE TO oml_user; GRANT CREATE VIEW TO oml_user; GRANT CREATE MINING MODEL TO oml_user; GRANT EXECUTE ON CTXSYS.CTX_DDL TO oml_user;READ or SELECT privileges are required for data that is not in your schema. For example, the following statement grants SELECT access to the sh.customers table. GRANT SELECT ON sh.customers TO oml_user; |
CopyRight 2018-2019 办公设备维修网 版权所有 豫ICP备15022753号-3 |