I notice all the new oracle advanced analytics 12c docs now display new data miner nodes for graphs, SQL & R scripts, etc. Does this require DB version 12c, a new version of SQLDeveloper (to be released) or both?
Oracle Data Miner GUI 4.0 connected to an 11gR2 Database + OAA Option adds the new Graph node (box plots, scatter, histogram and bar charts), SQL query node (run arbitrary SQL scripts and registered R scripts), other miscellaneous enhancements and bug fixes. ODMr4 running on 12c adds support for the new algorithms (Expectation Maximization, PCA, SVD), more integrated/automated Oracle Text integration/Text Mining, Prediction Details, Trees can mine text, tec. See latest OAA 12c presentation posted on OAA on OTN web site for more: http://www.oracle.com/technetwork/database/options/advanced-analytics/oaa12cpreso-1964644.pdf
First you need to download and install the Oracle Database EE which includes the Oracle Advanced Analytics/Oracle Data Mining Option--which installs by default. 12c is the latest version, but you can also run earlier versions. If not 12c, recommend 11gR2.0.3 Oracle Database Software Downloads</title><meta name="Title" content="Oracle Database Software D…
Then you need to download the latest version of SQL Developer which includes the Oracle Data Miner GUI extension and follow instructions for setting up ODM'r. Oracle SQL Developer Downloads</title><meta name="Title" content="Oracle SQL Developer Downloads… and http://apex.oracle.com/pls/apex/f?p=44785:24:3082890040080::NO:24:P24_CONTENT_ID%2CP24_PREV_PAGE:5272%2C2
There is a new SQLDEV/ODM'r 4.0 GUI that should be posted to OTN any day now as an Early Adopter 1 version. Watch the SQLDEV and ODM'r OTN web pages for updates.
The Oracle Advanced Analytics/Oracle Data Mining Option supports 15 hi-perf in DB SQL functions dm algorithims (clustering, classification, associations, etc.) but not forecasting e.g. ARIMA. Howvever, you can call R packages for time series via embedded R integration capabilities of OAA/ORE. Any aggregations of e.g. POS data for input to the forecasting models can be performed using the hi-perf SQL aggregation functions and then the summary results pushed to the R package and run via embedded mode. Hope this helps.