uriahf / performance_evaluation_of_prediction_models

Home Page:https://predictionperformance.netlify.app/performance_evaluation_of_prediction_models.html#/title-slide

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Cool intuition about Lift

uriahf opened this issue · comments

Thanks to Yatir Ben Shlomo:

One might consider Lift as the proportion of Sens/PPCR:

Allocating resources for 20% of the population (PPCR 0.2) will lead to Sens of 0.6 which is 3 times better than a random guess that will lead to Sens of 0.2 (= PPCR).

For random guess: PPCR = Sens, PPV = Prevalence.