A Survey of Artificial Neural Network-Based Modeling in Agroecology

摘要: Agroecological systems are difficult to model because of their high complexity and nonlinear dynamic behavior. The evolution such depends on a large number ill-defined processes that vary in time, whose relationships often highly non-linear very unknown. According Schultz et al. (2000), there two major problems when dealing with modeling agroecological processes. On the one hand, is an absence equipment able capture information accurate way, other hand lack knowledge about systems. Researchers thus required build-up models rich poor-data situations, by integrating different sources data, even if this data noisy, incomplete, imprecise.