Bowl Championship Series -- Analysis, Expert systems (Computer science), Bayesian statistical decision theory, Knowledge management
In college football the BCS (Bowl Championship Series) attempts to leverage both human experts and computer systems in an attempt to rank and identify the best teams. This is an example of trying to predict which teams are superior based on human knowledge and Bayesian statistics. Humans vote based on observations from the games, while computers use data and heuristics to try to identify the best teams. This domain can be used to observe and compare experts and computers moving from data to information to knowledge, and possibly gaining wisdom.
It may not be possible to know a priori which team will win any certain game, but a system should be able to determine the most likely teams to win or at least accurately rank the teams based on their success. The value of a ranking system should be measured against a posteriori knowledge. In some subject areas it may not be possible to collect this information, or there may not be enough data available to determine if experts or intelligent systems are making correct decisions or suggestions. Using a subject where human experts and computer systems make predictions on a weekly basis, and the results from those decisions are validated weekly, it is possible to evaluate the effectiveness of the system. These results could allow us to better understand what statistical data is important when a number of variables influence the results. Other domains, such as weather and economic forecasting, may be able to look at the BCS results in order to effectively implement optimization or ranking solutions.
Jensen, Deron Eugene, "The BCS: Evaluating Knowledge of Experts and Computer Systems" (2010). Engineering and Technology Management Student Projects. 870.