Publicação

Geosimulation and spatial analysis: linking Cellular Automata and Neural Networks to Forecast Land Use/Cover Change

Ver documento

Detalhes bibliográficos
Resumo:The geosimulation is an emergent field of inquiry that advocates the use of computational intensive methods of spatial analysis as the ones that appeal to heuristic search, neural nets and cellular automata. This work presents a method to simulate the land use/cover evolution in a rural/urban fringe reality, linking neural networks and cellular automata (CA) in a GIS environment. The simulation of such alterations appealing solely to cellular automata is not convenient, because these models, in its more conventional form, comprise limitations in the definition of the space parameters and the transition rules. In this work a neural net is used to survey the importance degree that each prediction variable (probability) has in the geographic constraints. These variables are gotten with resource to GIS.
Autores principais:Tenedório, José A.
Outros Autores:Rocha, Jorge; Morgado, Paulo; Encarnação, Sara
Assunto:Cellular automata Neural networks Land use/cover change
Ano:2005
País:Portugal
Tipo de documento:capítulo de livro
Tipo de acesso:acesso aberto
Instituição associada:Universidade de Lisboa
Idioma:inglês
Origem:Repositório da Universidade de Lisboa
Descrição
Resumo:The geosimulation is an emergent field of inquiry that advocates the use of computational intensive methods of spatial analysis as the ones that appeal to heuristic search, neural nets and cellular automata. This work presents a method to simulate the land use/cover evolution in a rural/urban fringe reality, linking neural networks and cellular automata (CA) in a GIS environment. The simulation of such alterations appealing solely to cellular automata is not convenient, because these models, in its more conventional form, comprise limitations in the definition of the space parameters and the transition rules. In this work a neural net is used to survey the importance degree that each prediction variable (probability) has in the geographic constraints. These variables are gotten with resource to GIS.