An Innovative Index for Evaluating Urban Vulnerability on Pandemic Using LambdaMART Algorithm

The COVID-19 pandemic has significantly changed urban life and increased attention been paid to the in discussions of vulnerability. There is a lack methods incorporate dynamic indicators such as vitality into evaluations In this research, we use machine learning establish an Pandemic Vulnerability Index (PVI) that measures city’s vulnerability takes important aspect this. proposed PVI constructed using 140 statistic variables 10 variables, data from 47 prefectures Japan. Factor Analysis used extract factors may affect city vulnerability, LambdaMART algorithm aggregate predict results show can relative seriousness two weeks with precision more than 0.71, which meaningful for taking controlling advance shaping society’s response. Further analysis revealed key affecting including size, transit station vitality, medical facilities, emphasizing precautions public transport systems new planning concepts compact city. This research explores application techniques indicator establishment incorporates assessments, contribute improvements assessments sustainable cities while facing challenges pandemic.