Table 1.
The performance of data mining algorithm given worst case/best case assumptions.
Table 2.
The performance of QOF criteria given worst case/best case assumptions.
Table 3.
The performance of expert knowledge driven algorithm given worst case/best case assumptions.
Fig 1.
RA phenotyping model: (Left) C5.0 decision tree. The value of a variable is the number of its occurrence. At each final node, RA = 1 indicates classification for positive RA and RA = 0 negative classified for RA. (Right) Decision rules generated from the C5.0 decision tree.
Table 4.
Algorithm code definitions.
Fig 2.
Prevalence of RA for 2000–2010 in the ABMU region, Wales, UK.