Skip to content

Statistical Learning for Overestimated Population Adjustments

Notifications You must be signed in to change notification settings

violetacln/SLOPA

Repository files navigation

SLOPA

Statistical Learning for Overestimated Population Adjustments

Here we share our workflows/R-scripts for using classifiers/ensembles and evaluating them, in order to predict the status of individuals (as present or absent) from a given population, provided a set of their attributes (demographic, social) are known. The earliest script established the exploratory and modeling steps necessary to fit and evaluate any such ML models, ensembles included. An additional code was focused on the random forest classifier and optimised for several performance measures.

Two papers presented at NTTS-2021 and NSM-2022 conferences are shared here as well as our paper in the Statistical Journal of the IAOS (2023). This is an open access paper and uses the code shown in the file above. It is the most general and it compares multiple ML models, it reports on their performance measures and associated uncertainty and chooses the optimum regimes, depending on the goal of the analysis.

About

Statistical Learning for Overestimated Population Adjustments

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages