We investigated the utility of integrating a geographic information system and a neural network for analyzing patterns in understory vegetation growth in a historic landscape. An artificial neural network was developed and trained to predict the tendency of honeysuckle to appear on a particular site, based on the biophysical characteristics of that site. The net was trained to associate patterns in biophysical characteristics at specific sample plots with the condition of honeysuckle growth found at those locations. Once trained on the data from the sample plots, the neural network recalled data fromthe entire site. The pattern in honeysuckle vegetation growth predicted by the neural net agreed fairly closely with field observations of the actual site. However, problems with overspecialization of the original neural network required a careful redevelopment of the training data.
|Original language||English (US)|
|Number of pages||6|
|State||Published - Dec 1 1997|
ASJC Scopus subject areas
- Computer Science(all)