I had the opportunity to attend the R/Finance 2012 Applied Finance with R conference two weekends ago at the University of Illinois, Chicago. The host of the event was the UIC International Center for Futures and Derivatives from the College of Business Administration, Gib Bassett, Director. I’d been to two earlier R/Finance conferences, the last in 2010. There were 200 participants then compared to in excess of 300 international attendees in 2012. Kind of a NATO Summit preview for Chicago.

I guess it’s a sign of the educational times, but the University of Washington Professional and Continuing Education division was an event sponsor this year, touting its online Master of Science in Computational Finance and Certificate in Financial Risk Management programs. I like the seeming diversity of the curricula, involving computer science, statistics, finance and economics. Washington’s highly regarded in each, so the $36K for the MS or $11K for the certificate program might well be good investments. Not surprisingly, the R programming environment with attendant finance packages is core to computational instruction. Faculty Doug Martin and Eric Zivot each presented some of his research.

The conference held sessions on May 11 and 12. There were 50 minute morning and afternoon keynotes each day, sandwiched around 20 minute feature presentations and batches of “lightening” talks. Many of the presentations were on financial topics showcasing R package solutions. Several, though, focused on general-purpose R software. The pace was furious for those who, like me, take notes. My head was spinning at the conclusion of the first series of lightening talks.

Legendary Blair Hull of Hull Investments/Ketchum Trading regaled the audience with 40 years of war stories in the Friday morning keynote. Hull started his “career” as a beat-the-dealer black jack player able to eke out a .008% advantage. Much of his talk revolved on explaining the 1987 crash, the flash crash and the “illusion of liquidity” associated with high frequency trading. Despite these anomalies, Hull’s no fan of SEC market regulations. I doubt he’ll be voting Democratic in November.

R analysts attempting to navigate the platform’s various optimization functions are often frustrated by the disparate interfaces to R’s linear, quadratic and non-linear packages. Presenter Stefan Theusbl and colleagues came to the rescue. They’ve developed an R Optimization Infrastructure (ROI) library which exposes a consistent object-oriented interface to these and other commercial optimizers, greatly standardizing their use. Kind of reminds me of the Zelig package for predictive models and the caret library for machine learning.

Friday afternoon keynoter Rob McCulloch addresed “Cholesky Stochastic Volatility.” Yikes. Once I settled back in my seat, I realized I could actually understand some of his highly-entertaining presentation. McCulloch’s an unabashed Bayesian who uses state space models to estimate dynamic portfolio covariance matrices – there, I said it. His biggest challenges? Choice of a prior distribution and performance hurdles that require parallel computing.

European academic and investment practitioner Kris Boudt presented on the important topic of investment performance measurement, characterized by Bruce Feibel as ”The calculation of portfolio return, risk and derived statistics stemming from the periodic change in market value of portfolio positions and transactions made into and within a portfolio, for use in the evaluation of historical fund or manager performance.” Boudt’s analysis deploys data from the Hedge Fund Research database to evaluate fund performance from 2005-11. He categorizes portfolios as equal, over or under performers based on a methodology that compares Sharpe ratios and adjusts results for random false discoveries. His overall finding? There aren’t many overperforming hedge funds. Buyers beware.

Being a slave to a daughter’s schedule to help with Mothers’ Day shopping kept me from the conference Saturday, much to my chagrin. I was especially anticipating several of the presentations more directly relevant to BI and predictive analytics, including “Lasso-Assisted, Vector Autoregression,” “News Sentiment Analysis Using R to Predict Stock Market Data,” “Visualizing Large Data with R,” “Large-Scale, Computationally Intensive Forecasting in R,” and “CppBugs: Native MCMC for R.” I plan to review the slide decks if they’re made available on the R/Finance website.

Kudos to R/Finance committee members Gib Bassett, Peter Carl, Dirk Eddelbuettel, Brian Peterson, Dale Rosenthal, Jeffrey Ryan and Joshua Ulrich for another terrific conference. The trajectory of R/Finance is nothing but positive. Indeed, last week I felt compelled to ask Gib about logistics for future conferences, given the tremendous growth over the last four years. How would R/Finance and UIC accommodate the heightened demand? Ever the economist, Gib deadpanned “charge more!”