Accounting for Dependence in a Flexible Multivariate Receptor Model

摘要: Formulation and evaluation of environmental policy depends on receptor models that are used to assess the number nature pollution sources affecting air quality for a region interest. Different approaches have been developed situations in which no information is available about these (e.g., exploratory factor analysis) composition assumed known regression measurement error models). We propose flexible approach fitting model when only partial source available. The use latent variable modeling allows direct incorporation subject matter knowledge into model, including physical constraints obtained from laboratory measurements or past studies. Because data often exhibit temporal and/or spatial dependence, we consider importance accounting such correlation estimating parameters making...