A COMPARISON BETWEEN PARAMETRIC AND NON-PARAMETRIC APPROACHES TO THE ANALYSIS OF REPLICATED SPATIAL POINT PATTERNS

摘要: The paper compares non-parametric (design-based) and parametric (model-based) approaches to the analysis of data in form replicated spatial point patterns two or more experimental groups. Basic questions for this kind concern estimating properties underlying process within each group, comparing between A approach, building on work by Diggle et. al. (1991), summarizes pattern an estimate reduced second moment measure K-function (Ripley (1977)) mean K-functions groups using a bootstrap testing procedure. approach fits particular classes model data, uses parameter estimates as summaries tests differences with without assumption common values across discusses how either can be implemented specific context single-factor experiment simulations show efficient when assumptions hold, but potentially misleading otherwise.