Calibrating cellular automata for urban development modelling using principal component analysis

Principal component analysis (PCA) is a powerful technique for extracting structure from high-dimensional datasets. In this paper, a PCA based cellular automata (CA) model for modelling urban development is presented. Compared to the conventional method of retrieving CA transition rules, the PCA model needs a small number of principal components to account for most of the structure in the datasets due to the noise reduction. The PCA-CA model is successfully applied in a fast growing area of Shanghai, eastern China. The results produced by the PCA-CA model shows that it matches well with the actual development of the case study area with relatively high accuracy.