Bayesian capture-recapture in social justice research
Also appears in collection : Jean-Morlet chair - Workshop: Young Bayesians and big data for social good / Chaire Jean-Morlet - Workshop : Jeunes Bayésiens et big data pour le bien social
Capture-Recapture (RC) methodology provides a way to estimate the size of a population from multiple, independent samples. While the was developed more than a century ago to count animal populations, it has only recently become important in Data For Social Good. The large number of samples with varying amounts of intersection and developed over a period of time, so often found in Data For Social Good projects, can greatly complicate conventional RC methodology. These conditions are ideal, however, for Bayesian Capture Recapture. This presentation describes the use of Bayesian Capture Recapture to estimate populations in Data for Social Good. Examples illustrating this method include new work by the author in estimating numbers of human trafficking victims and in estimating the size of hate groups from the analysis of hate speech in social media.
![$k$-sum free sets in $[0,1]$](/media/cache/video_light/uploads/video/2015-09-10_de_Roton-video--19cf70874d3c5af7378c268cab865b06.jpg)
