Data Mining, Machine Learning and Spatial Data Infrastructures for Scenario Modelling

Neil Sang and Mathew Aitkenhead

fig9The approaches discussed in relation to most scenario models relate to some form of simulation, mathematical model or representation in a formal modelling language. The range of computational or technological complexity involved is variable, but in most cases a very high degree of domain knowledge is also required with respect to the system under investigation.

Retaining the necessary skills on a wide range of problems is difficult, and indeed current scientific knowledge is not always sufficient to understand and represent a particular system. For large coupled systems, with a wide range of socio-economic, ecological and bio-physical systems interacting, this may not be feasible. For simpler systems, in specific cases, it may not be necessary. This chapter explores data mining as a pragmatic alternative or complementary approach. It also looks at the role of a spatial data infrastructure in lowering the cost of modelling work.

Back to Summary / Visit