BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20250504T230009EDT-4871i5Vvav@132.216.98.100 DTSTAMP:20250505T030009Z DESCRIPTION:Prasad Patil\, PhD\n\nAssistant Professor\n Department of Biosta tistics |\n Boston University School of Public Health\n\nWHEN: Wednesday\, January 29\, 2025\, from 3:30 to 4:30 p.m.\n WHERE: Hybrid | 2001 51³Ô¹ÏÍøCo llege Avenue\, Room 1140\; Zoom\n NOTE: Prasad Patil will be presenting fro m Boston\n\nAbstract\n\nAtmospheric concentration of ultrafine particles ( UFP\; particles <100nm) has the potential to affect human health in a dist inct manner from larger particles such as PM2.5. In a study of aviation-at tributable UFP from flight activity at Boston Logan Airport\, our team imp lements stationary monitoring at varying distances along flightpaths from the airport. One goal is to develop a predictive model for UFP using meteo rological and air traffic measurements that is generalizable to regions wh ere direct monitoring cannot be conducted. Due to variations in orientatio n to the airport and local environmental factors\, UFP measurement distrib utions from our monitoring sites are heterogeneous. There are advantages t o training prediction models on the combination of all site data or by wei ghted combinations of predictions from single-site models. We show that fo r unbiased models\, an optimal blending of predictions from both approache s performs at least as well as each scheme adaptively at differing levels of heterogeneity. We apply insights from these findings to propose an empi rical weighting strategy based on cross-site model performance and agreeme nt.\n\nSpeaker Bio\n\nPrasad Patil is an Assistant Professor of Biostatist ics at the Boston University School of Public Health. His primary research interests are reproducibility and replicability\, with a focus on machine learning and prediction in public health. Areas of application include ge nomic biomarker development\, air pollution monitoring\, opioid overdose s urveillance\, and replication of published scientific results. He complete d his PhD in biostatistics at Johns Hopkins BSPH and his postdoc at Harvar d CSPH/Dana-Farber Cancer Institute.\n DTSTART:20250129T203000Z DTEND:20250129T213000Z SUMMARY:Cross-site blending for generalizable prediction of UFP concentrati ons URL:/epi-biostat-occh/channels/event/cross-site-blendi ng-generalizable-prediction-ufp-concentrations-362676 END:VEVENT END:VCALENDAR