paperKB
coga / coga-kb
Help
Sign in

Chunk #67 — Discussion — Outlook

Source
Random forest versus logistic regression: a large-scale benchmark experiment.
Embedded
yes

Text

Before further studies are performed on tuning strategies, we insist that, whenever performed in applications of RF, parameter tuning should ideally always be reported clearly including all technical details either in the main or in its supplementary materials. Furthermore, the uncertainty regarding the “best tuning strategy” should in no circumstances be exploited for conscious or subconscious “fishing for significance”.