Grammars provide a means of generating large amounts of variability from relatively small resources, and seem especially relevant when the variation to be modelled is discontinuous since they describe the organisation symbols from some alphabet. In order to use grammars, it is desirable that they be inferred from observational data, and we present three examples of such inference. Two involve temporal successions, while the third is based on spatial transects of aquatic vegetation in wetlands in Victoria, Australia. The temporal successions were adequately represented by simpler regular grammars, suggesting that the processes involved may be modelled in a relatively simple manner. The spatial data requires a context-free grammar, reflecting the greater complexity of the spatial series, the uniqueness of each site, and the difficulties of modelling parallel systems with serial models. Examining the grammars further leads to some suggestions concerning the features of them which are important for ecological interpretation. It further leads to a novel view of the nature of groups in classification.
rs means book pdf free 15
2ff7e9595c
Comments