A variety of statistical techniques are suggested for the analysis of data where the incidence of respiratory cancer is thought to be related to a putative pollution source. Non-parametric kernel regression techniques and more formal point process models are considered. Parameter estimation based on maximum likelihood is discussed. Extensions to an existing model are proposed, and a new hybrid model is also fitted.
|Number of pages||14|
|Journal||Journal of the Royal Statistical Society Series A - Statistics in Society|
|Publication status||Published - 1994|