Poster Title
Comparing Statistical Methods for Predicting Human Behavior: An Example with Team Performance
Faculty Sponsor, if applicable
Goran Kuljanin
Project Abstract
This research developed and compared statistical methods for predicting human behavior. We investigated the efficacy of four statistical methods for predicting team performance of professional basketball teams. Each statistical method was evaluated on six accuracy metrics. The moving average and Monte Carlo methods performed the best across all six accuracy metrics. The results suggest that an effective method for predicting team performance involves computing updated predictor values as new dynamic information is gathered on teams.
Type of Research
Doctoral-Undergraduate Opportunity for Scholarship (DUOS)
Preview
Presentation Year
May 2017