Soft Computing Approach to Imperfection Propagation: Application to Land Cover Change Prediction
Résumé
Land cover change prediction is an important issue for several fields such as urban
sprawl prevention, planting status of agricultural products, and supervision of desertification and
erosion. The process of change prediction is usually characterized by several types of
imperfection. Most works in literature attempt to resolve this problem by developing or
improving models that take into account imperfection related to data. However, these works
disregard the imperfection related to the input of their models and its propagation through their
models. This paper proposes an approach that propagates imperfections through a land cover
change model. It allows estimating the imperfection in the output of land cover change model
from the imperfection in the inputs. This helps us to identify robust conclusions allowing
remotely sensed users to make proactive and knowledge-driven decisions. The proposed
approach is validated by using a model allowing the prediction of urban changes.