Abstract
Agricultural land-suitability assessment involves the analysis of a large variety and amount of physiographic data. Geographical information systems (GISs) may facilitate suitability assessment in data collection. To generate accurate results from the data, appropriate suitability-assessment methods are required. However, the assessment methods which can currently be used with GISs, such as that developed by the United Nations Food and Agriculture Organization and the statistical pattern—classification method, have limitations which may lead to inaccurate assessment. An artificial neural network is an effective tool for pattern analysis. A neural network allows decision rules of greater complexity to be applied in pattern classification. By formulating the land-suitability-assessment problem into a pattern—classification problem, neural networks can be used to achieve results of greater accuracy. In this paper, a neural-network-based method for land-suitability assessment is discussed, and a set of neural networks is described. The integration between the neural networks and a GIS is addressed, and some experimental results are presented and analyzed.
Get full access to this article
View all access options for this article.