We’re pleased to announce the latest release of Oracle R Enterprise, now available for download. Oracle R Enterprise 1.3 features new predictive analytics interfaces for in-database model building and scoring, support for in-database sampling and partitioning techniques, and transparent support for Oracle DATE and TIMESTAMP data types to facilitate data preparation for time series analysis and forecasting. Oracle R Enterprise further enables transparent access to Oracle Database tables from R by enabling integer indexing and ensuring consistent ordering between data in R data frames and Oracle Database tables. The latest release also includes improved programming efficiencies and performance improvements.
The key additions in version 1.3 include:
Enhanced Model Scoring: The new package OREpredict enables in-database scoring of R-generated models. Supported models include linear regression (lm) and generalized linear models (glm), hierarchical clustering (hclust), k-means clustering (kmeans), multinomial log-linear models (multinom), neural networks (nnet), and recursive partitioning and regression trees (rpart).
Oracle Data Mining Support: The new package OREdm provides an R interface for in-database Oracle Data Mining predictive analytics and data mining algorithms. Supported models include attribute importance, decision trees, generalized linear models, k-means clustering, naive bayes and support vector machines.
Neural Network Modeling: A new feed-forward neural network algorithm with in-database execution.
Date and Time Support: Support for Oracle DATE and TIMESTAMP data types and analytic capabilities that allow date arithmetic, aggregations, percentile calculations and moving window calculations for in-database execution.
Sampling Methods: Enables in-database sampling and partitioning techniques for use against database-resident data. Techniques include simple random sampling, systematic sampling, stratified sampling, cluster sampling, quota sampling and accidental sampling.
Object Persistence: New capabilities for saving and restoring R objects in an Oracle Database “datastore”, which supports not only in-database persistence of R objects, but the ability to easily pass any type of R objects to embedded R execution functions.
Database Auto-Connection: New functionality for automatically establishing database connectivity using contextual credentials inside embedded R scripts, allowing convenient and secure connections to Oracle Database.
When used in conjunction with Oracle Exadata Database Machine and Oracle Big Data Appliance, Oracle R Enterprise and Oracle R Connector for Hadoop provide a full set of engineered systems to access and analyze big data. With Oracle R Enterprise, IT organizations can rapidly deploy advanced analytical solutions, while providing the knowledge to act on critical decisions.