Inferring Infection Transmission Parameters That Influence Water Treatment Decisions
针对美国环保署评估饮用水系统微生物感染风险的需求，开发了基于随机过程的推断方法，利用地方性数据估计影响水处理政策的关键传播参数，并以纽约市数据为例验证。
One charge of the United States Environmental Protection Agency is to study the risk of infection for microbial agents that can be disseminated through drinking water systems, and to recommend water treatment policy to counter that risk. Recently proposed dynamical system models quantify indirect risks due to secondary transmission, in addition to primary infection risk from the water supply considered by standard assessments. Unfortunately, key parameters that influence water treatment policy are unknown, in part because of lack of data and effective inference methods. This paper develops inference methods for those parameters by using stochastic process models to better incorporate infection dynamics into the inference process. Our use of endemic data provides an alternative to waiting for, identifying, and measuring an outbreak. Data both from simulations and from New York City illustrate the approach.