Thesis (M.Eng.) - Asian Institute of Technology, 1996

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering.

Artificial Neural Networks are new technologies for classifications. They are able to process incomplete and imprecise data and to detect non-linear relations in the data. Artificial learning algorithms can be subdivided into two types, supervised and unsupervised. Neural networks learn in massively parallel and self-organizing way. Unsupervised learning neural networks, like Kohonen's self-organizing feature maps (Kohonen, 1989), learn the structure of high-dimensional data by mapping it on low-dimensional topologies, preserving the distribution and topology of the data. In this thesis the Kohonen self-organizing feature map is applied to classification of a land cover data set. The data was collected from existing database of land cover regions in Cambodia that were · expertly labeled into many classes. Rule extraction extracts land cover classes produced by self-organizing methods for the queries of knowledge. However, a rule generation algorithm of rule extraction out of the neural network, which could be used by the geological expert.