From June 17–28, 2013, the IMA will hold its New Directions Short Course, entitled “Applied Statistics and Machine Learning.” Organized by Bin Yu (University of California, Berkeley) and David Madigan (Columbia University), the course provides opportunities for establishing collaborations, building connections, and possibly shifting one’s research focus. Past attendees have cited the course as the impetus for publishing papers and books, beginning new projects, and joining forces with others in the field to form new, inspired partnerships. Because statistics as a discipline exists to extract meaningful information from data, statistics and statisticians now play an ever more central role in a data-driven world. Applied statistics aims to answering domain questions by collecting and analyzing data through statistical critical thinking and statistical methodology. Machine learning is a recent, thriving field of statistics that explicitly takes computation into account and its researchers affiliate with departments of computer science, statistics, electrical engineering, and mathematics.
This two-week course will introduce participants to a broad array of modern statistical concepts and techniques with a focus on critical thinking and practical data analysis. The statistical software R will be used extensively and students are expected to have at least rudimentary knowledge of R prior to the course. The course will cover exploratory data analysis (visualization, dimension reduction, clustering), statistical modeling (linear models, generalized linear models, logistic regression, graphical models), and statistical computation (Monte Carlo, Markov chain Monte Carlo, convex optimization). The course will also cover regularized and large-scale modeling techniques. Both frequentist and Bayesian perspectives will be considered. Specific applications include inference from large-scale observational healthcare data, localization in wireless networks, fMRI brain signals evoked by natural stimuli, remote sensing, and text analysis.
More information and an online application are available online at http://www.ima.umn.edu/2012-2013/ND6.17-28.13. The intended audience is mathematical scientists with interests in data analysis but with a limited statistical background. We assume participants will have had an introductory course in statistics and will be familiar with elementary statistical concepts, such as sampling, confidence intervals, and hypothesis tests. We will make extensive use of R and will provide tutorial materials in advance for participants that are not familiar with basic R.
The applications process is now open; applications are due April 15, 2013.
(Photo courtesy of http://thehandsweshake.com.)