Variable selection methods for multiple regressions influence the parsimony of risk prediction models for cardiac surgery
A surprisingly readable Australasian manuscript on risk prediction models for cardiac surgery. Interesting that the class of dyspnoea was not captured in 5.28% of the operated patients.
0
0 comments