利用人工智能发现差距并改变低收入和中等收入国家的医疗保健.pdf

1、Data for GoodChanging the Shape of Medical Deserts to Save LivesActionable Data Initiative How can an improvement in efficiency save millions of lives?By using data science to create a more efficient marketplace in global health philanthropyMarket dynamics for global healthcare delivery are brokenTh

2、e global healthcare landscape is plagued by information asymmetry,creating an inefficient“lemon market.”As a result,resources are not reaching the key actors who need them most.CHALLENGEMedical ProfessionalsGovernmentsNonprofit Organizations143 Million People are waiting for surgery in Lower and Mid

3、dle Income countries.RESULTSLeveraged Big Data and ML to generate the largest database of healthcare facilities and nonprofits across 72 low and low-middle income countriesFOUNDATIONAL WORKPublished results with Taylor&Francis in first World Compendium of Healthcare Facilities and Nonprofit Organiza

4、tionsBOOK PUBLICATIONIntegrated book with a digital counterpart to further explore the database and provide donation and volunteer opportunities.BOOK PUBLICATIONBuilt VFMatch.org,first digital marketplace for global health philanthropyDIGITAL PLATFORMA trusted and easy-to-use tool for both supply an

5、d demand side participantsBeta Available TodayTrain an ML model which scrapes websites in real-time and predicts nonprofit websitesCREATING THE NGO DATASETNonprofit WebsitesCommercial WebsitesHealthcare FacilitiesMachine Learning TrainingInput Data:Tens of Thousands of Websites Indexed ML ModelNew U

6、nknown WebsitePrediction NonprofitHospitalCommercialConservatively exclude facilities with a building footprint too small to be a hospitalRun hospital name against multilingual list of keywords to exclude non-hospitals(e.g.“dispensary”,“clinic”,“centre de sant”etc.)Filter out candidate facilities wi