High performance computing with R (various topics revisited and updated)

*** New venue!!! Read below ***

High performance computing (HPC) with R

by Ryan Rosario

Abstract:

Over the years, several attempts have been made to bring parallel processing to the R environment. HPC in R is a large topic, and all are encouraged to share their experiences. This talk will serve as an update to my previous talk about high performance computing and parallelization in R. I will introduce the new "parallel" library in R, which combines both the "snow" and "multicore" packages. I will also talk about more recent advances in using R with Hadoop via the "rmr" package.

Speaker bio:

Ryan Rosario is Chief Data Scientist/Research Engineer at GumGum, an in-image ad network in Santa Monica. He is also a Ph.D. Candidate at UCLA Department of Statistics and holds an M.S. in Computer Science from UCLA. Ryan uses open-source software extensively in his work and research (Python, Hadoop and friends, NoSQL) and works with these tools both locally and in the cloud (EC2). He uses R for all of his data analysis, and for prototyping machine learning algorithms and models. His interests include text mining, natural language processing, machine learning, graph/network analysis and data mining.

Please RSVP as places are limited.

New venue: Shopzilla will host this event (and kindly provide pizza and drinks). The room for the meeting is on the 4th floor.

Parking: As I understood, garage parking will be partially validated, you'll have to pay $3 cash (otherwise garage parking is $12, cash only!). Some street parking is also available nearby (if you find a spot, and I'm not sure about the price, but most likely free after 6pm). Please consider your options in advance and please do not flood the mailing list with complaints after.