Oracle Machine Learning for SQL (OML4SQL) on Amazon RDS for Oracle enables building, training, and scoring ML models entirely within the database using SQL and PL/SQL — no data movement, no external ML platforms required. The walkthrough covers setting up private VPC infrastructure, configuring database users with required grants, training a credit risk classification model using the DBMS_DATA_MINING package with AutoML (automatic algorithm selection), and scoring predictions using standard SQL PREDICTION() functions. The guide covers both Oracle Database 26ai (CDB architecture) and Oracle Database 19c (non-CDB), noting that OML4SQL supports classical algorithms like Random Forest, Naive Bayes, Decision Tree, and SVM. Key caveats include Enterprise Edition licensing requirements, lack of native model versioning, and unsuitability for deep learning or Python-based workflows.
