Wednesday Dec 26, 2012

Oracle R Enterprise 1.3 released

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.

Stay tuned for blogs about the new ORE 1.3 features in upcoming posts. You can find more details about the features in Oracle R Enterprise 1.3 in our New Features Guide and Reference Manual.

About

The place for best practices, tips, and tricks for applying Oracle R Enterprise, Oracle R Distribution, ROracle, and Oracle R Advanced Analytics for Hadoop in both traditional and Big Data environments.

Search

Archives
« December 2012 »
SunMonTueWedThuFriSat
      
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
27
28
29
30
31
     
Today