Figure Skating Judgement: To Combine or Not To Combine

Faculty Mentor: Frank Hsu (Fordham)

Description

Sports competition judgement, ranking books, movies, or candidates as well as prediction of election outcomes all involve various scoring systems by multiple judges, experts, models or systems. One of the many challenges is how to combine these multiple scoring systems to reach a final scoring system (or ranking system) which is fair, accurate, and effective. This project explores figure skating competition judgement using machine learning techniques and combinatorial fusion algorithms. Both score and rank combinations are considered and participants will learn the pros and cons of each combination and apply this insight to their everyday life and future career.

Objectives and Learning Goals

Participants, after completing the project, will have achieved the following learning and research objectives:

  • Learned the basic concepts of scoring and ranking in everyday life w.r.t. competition, judgement, prediction or decision making.
  • Learned the fundamental difference between scoring and ranking items in a data set and the diversity between multiple judges, experts, models, systems, and algorithms.
  • Learned how to best form a group (or committee) of judges, experts or professionals as well as algorithms, systems, softwares or models.
  • Learned how to catch a cheater in a competition judgement or an intruder in information and cyber security.