To register for this course, please go to http://www.odum.unc.edu/datamatters

Predictive analysis exploits patterns found in transactional and other data to identify risks and opportunities. It is used in actuarial science, marketing, financial services, telecommunications, retail, travel, healthcare, pharmaceuticals and other fields. This course will discuss a number of statistical and machine learning methods that have been used for forecasting social and economic time series, that is, a sequence of data points measured typically at successive points at uniform time intervals. The statistical methods will focus on classical time series analysis, with particular attention to generating robust models. The machine learning methods will focus on how a variety of “big data” approaches originally developed for cross-sectional analysis can be adapted to forecasting. While the course will primarily focus on methodology, it will also consider the work of Daniel Kahneman, Philip Tetlock, Nassim Taleb, and others on the difficulties in qualitative forecasting, as well as limits to forecast accuracy.

If you have questions, please contact Jill_Stevens@unc.edu.

**Cancellation Policy The following refund policy will apply in the event of cancellations:

• Full refunds, less a 33% cancellation fee, will be issued for any course enrollments canceled by you before May 14, 2014.

• Full refunds, less a 50% cancellation fee, will be issued for any course enrollments canceled by you between May 14, 2014 and June 15, 2014.

• Full refunds, less a 75% cancellation fee, will be issued for any course enrollments canceled by you after June 15, 2014.

• Prior to the event start date, if we cannot accommodate you in a course you registered for, we can issue a credit toward any UNC Odum Institute course (except for ICPSR courses) or issue you a full refund for the course. Credits must be used by September 30, 2014.