Aster-R: Become an R power user by harnessing the muscles of Aster

Learn Data Science
Teradata Employee

Please check out the attached PDF document from Dr. Diego Klabjan's 2015 Teradata Partners Presentation.  The attached documentation also includes source code and examples.


Diego Klabjan

Partner, Opex Analytics

Professor, Northwestern University

Director, Master of Science in Analytics

In this PDF document he covers the business case of R and AsterR

•R – preferred tool of many data scientists


–Link between R and the power of Aster

–From R environment

•Access distributed Aster functions via SQL-H

•Draw parallels between R and Aster-R

AsterR Strengths:

•Natural fit with Native R in terms of environment and general syntax

•Facilitates text cleansing and parsing

–Hard task in R

•Intuitive function and option names

•Easy to pick up where one last left off because of table creation

•In terms of functions, easier to use than Native R

He also demonstrates how to build a Cross Validation Function in R


•One round of cross-validation

–Partition a sample of data into complementary (equal size) subsets

–Perform the analysis on one subset

»Called the training set

–Validate the analysis on the other subset

»Called the validation set (or test set)

•Further reduce variability

–Conduct multiple rounds of cross-validation

»Using different partitions and averaging all results