|
Abstract:
|
When a drinking water treatment facility is in violation, this non-compliance should be addressed with an appropriate enforcement action by the governing health and environmental agency. In this case, the response time for a violation is critical in that a prompt response to a violation by the agency is necessary in order to minimize potential public health risk. We estimate the response time using Bayesian survival models. The models considered are parametric Weibull model, non-parametric Cox regression model using the counting process, and semi-parametric Accelerated Failure Time model using the Mixed Dirichlet process. Estimates from the recent violation data using these methods were remarkably comparable.
|
