Uninformed: Informative Information for the Uninformed

Vol 5» 2006.Sep


BayesCompare

Considering the significant disadvantage that BayesCompare is at relative to the other metrics, it performed quite well. It is quite possible that in practice BayesCompare could be the most accurate. This could be accomplished by mapping the probability of seeing particular chipset, device-driver implementation back to the marketshare of the chipset. This optimization is not implemented in the current system, and both flavors of BayesCompare do worse than the other metrics presented.