In view of the restricted system knowledge and data availability as well as the tremendous spatial and temporal heterogeneity of the real processes concerned, the choice of suitable modelling approaches in agroecology is not a trivial problem. For most problems different modelling approaches are equally accurate and reliable, and they can coexist.

If you wish to purchase Mathematica for Students see Accessing. Simulating Neural Networks with Mathematica. You can download the Mathematica notebook files below to view with Mathematica or Wolfram CDF Player (which is free). IEEE 87(9), 1423–1447 (1999) S. Samarasinghe, Optimum Structure of Feed Forward Neural Networks by SOM Clustering of Neuron Activations. Proceedings of the International Modelling and Simulation Congress (MODSM) (2007) Neural Networks for Mathematica, (Wolfram Research, Inc. USA, 2002) J. Sietsma, R.J.F..

In situations where the diversity of data is great and relationships between cause and effects are vague, neural networks seem to be a promising tool. Opportunities and limitations of the neural network paradigm in agroecological modelling are critically discussed. Two neural network applications, in habitat quality and biomass growth modelling, exemplify different application aspects. To improve the applicability of neural networks, input data pre-processing and the combination of networks with other AI techniques are advocated.

• Previous article in issue • Next article in issue. Dell Gx270 Usb Drivers.