METHODS AND ALGORITHMS OF THE NEURONET CLASSIFICATION AND PROTOTYPES DEFINITION IN GIS

摘要：

Geographical models of specific geosystems, namely the complex hierarchical systems with the spatial-temporal organization, are used to study and forecast the processes occurring in an environment. The major stage of modelling of such systems is classification of spatial geotypcs and definition of functional dependence of process development in time. For these purposes the mechanism of neural networks, having classifying and generalizing properties, is offered. The use of algorithms of classification and clusterization for modelling of geosystems is examined using the examples of (1) forecasting a climatic situation as the process developing in time and (2) the classifying description of forms of the topography reflecting the spatial condition of a geosystem.

作者: I.V.Starodub T.P.Varshanina O.A.Plisenko

作者单位: Adyghe Stale University Centre of GIS technologies Universitelskaya St., 208, Maykop 385000, Russia

会议类型: 国际会议

会议名称: 第十四届国际地图与地理信息系统学术研讨会

会议地点: 乌鲁木齐

会议语种:英文

页码: 515-519

在线出版日期: 2008-06-29（万方平台首次上网日期，不代表论文的发表时间）