Oracle Data Miner 4.2 (part of SQL Developer 4.2) got released as an Early Adopter versions (EA) a few weeks ago.
With the new/updated Oracle Data Miner (ODMr) there are a number of new features. These can be categories as 1) features all ODMr users can use now, 2) New features that are only usable when using Oracle 12.2 Database, and 3) Updates to existing algorithms that have been exposed via the ODMr tool.
The following is a round up of the main new features you can enjoy as part of ODMr 4.2 (mainly covering points 1 and 2 above)
- You can now schedule workflows to run based on a defined schedule
- Support for additional data types (RAW, ROWID, UROWID, URITYPE)
- Better support for processing JSON data in the JSON Query node
- Additional insights are displayed as part of the Model Details View
- Additional alert monitoring and reporting
- Better support for processing in-memory data
- A new R Model node that allows you to include in-database ORE user defined R function to support model build, model testing and applying of new model.
- New Explicit Semantic Analysis node (Explicit Feature Extraction)
- New Feature Compare and Test nodes
- New workflow status profiling perfoance improvements
- Refresh the input data definition in nodes
- Attribute Filter node now allows for unsupervised attribute importance ranking
- The ability to build Partitioned data mining models
Look out for the blog posts on most of these new features over the coming months.
WARNING: Most of these new features requires an Oracle 12.2 Database.