ANNIE™ MOORE World’s first refugee matching software Based on: Trapp, Teytelboym, Martinello, Andersson and Ahani (2018)
Matches are suggestive and HIAS retains complete control of process Employment prospects of Past employment data any refugee in Advanced any affiliate machine learning Best overall and econometrics employment outcomes for all resettled State-of-the-art integer refugees! optimization Match Refugee data and software needs of information refugees & provided by affiliate affiliates services
Affiliate Mismatch Information
Performance of Annie • Accuracy* : 80% – Satisfies all integration needs constraints : language, nationality, large family, single parent … – Employment gain is over 30% above observed baseline in 2017. – Almost every refugee’s employment probability goes up. – With minimum average family size constraints: 25% gain .
Directions for improvement • Using other integration outcomes • Incorporating greater diversity • Dynamic quota management • Involving new affiliates • Preferences
Thank you! Annie™ M atching and O ptimizing O utcomes for R efugee E mpowerment Annie Moore, 1892
Recommend
More recommend