AStatistical Comparison ofGrammatical Evolution Strategies intheDomainof HumanGenetics
Bill C. White, Joshua C. Gilbert, Jason H. Moore · 2005
Detecting andcharacterizing genetic predic- tors ofhumandisease susceptibility isanimportant goal inhumangenetics. Newchip-based technologies are available thatfacilitate themeasurement ofthousands ofDNAsequence variations across thehumangenome. Biologically-inspired stochastic search algorithms are expected toplay animportant roleintheanalysis of these high-dimensional datasets. Wesimulated datasets withupto6000attributes using twodifferent genetic models andstatistically compared theperformance of grammatical evolution, grammatical swarm,andran- domsearch forbuilding symbolic discriminant func- tions. Wefound nostatistical difference amongsearch algorithms within this specific domain.