For many years, databases were mainly used to store and retrieve data. Today, with the rise of Artificial Intelligence (AI), databases are expected to do much more. Oracle Database is evolving to meet this demand by becoming an AI ready data platform, where data storage, analytics, and AI capabilities are incorporated together.
Traditional databases required data to be moved out to external tools for machine learning and AI processing. This often caused performance issues, higher costs, and security concerns. Oracle Database changes this by allowing AI and machine learning to run directly inside the database, close to the data, which is known as "in-database machine learning".
Oracle Machine Learning (OML) provides built-in
machine learning algorithms. These algorithms can be used with SQL, making it
easy for database developers and analysts to adopt AI without learning complex
new tools.
Below is a simple example.
Step 1: Creating the Model
BEGIN
DBMS_DATA_MINING.CREATE_MODEL(
model_name => 'test_model',
mining_function =>
DBMS_DATA_MINING.CLASSIFICATION,
data_table_name =>
'test_data',
case_id_column_name => 'test_id',
target_column_name => 'test_c',
settings_table_name => 'mod_settings'
);
END;
/
Step 2: Model Parameters & Settings Table
INSERT INTO mod_settings VALUES (DBMS_DATA_MINING.ALGO_NAME, DBMS_DATA_MINING.ALGO_DECISION_TREE);
INSERT INTO mod_settings VALUES (DBMS_DATA_MINING.PREP_AUTO, DBMS_DATA_MINING.PREP_AUTO_ON);
Step 3: Using the Model for Prediction
SELECT test_id, PREDICTION(test_model USING *) AS test_prediction FROM test_data;
With just SQL, users can build and apply machine learning
models without moving data outside the database.
Oracle Database is also evolving to support Generative AI use cases. With vector data types and vector search, Oracle can store embeddings and perform semantic searches.
Running AI inside the database improves performance and reduces latency. Since data does not need to be copied to external systems, sensitive information remains protected.
Oracle Database also provides strong
security, access control, and governance, which are critical for enterprise AI
and responsible AI practices.
Comments
Post a Comment