On May 7 and 8 there will be a Conference on Predictive Inference and its Applications at Iowa State University. Predictive inference, as best as I understand it, is different from more traditional views of statistical inference in that while most inference seeks to predict model parameters, which are then used to predict the future, predictive inference seeks to use previous observations to predict future observations. In that respect predictive inference is closer to the broader discipline of forecasting.
The speaker list is set and will include a number of sports-related talks. One talk, by Stephanie Kovalchik of Victoria University and Tennis Australia on predicting tennis player emotions from a single-camera video feed, was a poster finalist at the most recent Sloan Sports Analytics Conference. Another sports-related talk, by David Harville of Iowa State, covers model-based prediction applied toward predicting outcomes of college football games. A third talk, by Carl Morris of Harvard, is ambitiously and broadly titled “Prediction in Sports”.
If you can get to Ames during the first week of next month it might be worth your time to attend the conference. There is also a poster session open to students and emerging researchers for which abstracts are due April 15. If you’re willing to stay an extra day, there will be an inaugural Midwest Statistical Machine Learning Colloquium that also appears to be worth attending.
Thanks to the Statistics in Sports section of the American Statistical Association for bringing this conference to my attention.