Integrating machine learning with Markov chain and cellular automata models for modelling urban land use change

Modelling urban land use change is of profound concern to environmental scientists who have found cellular automata models very attractive for simulating dynamics. The quest suitable predictive improve realistic simulation has resulted in the several notable calibrations. Cellular model become and one strongest growth due its simplicity possibility evolution. However, inability include driving forces process warranted further calibrations minimize this weakness. To address problem, contrary previous calibrations, research presents a novel integration support vector machine, Markov chain modelling. Support machine introduced as learning technique mine impact explanatory variables that drive change. employed transition probabilities between various epochs while are used implement incremental discrete time steps based on neighbourhood interaction from an initial future time. This modelling implemented using Landsat data acquired 1984, 2000 2015 over Lagos Nigeria, Africa’s most populous city. Urban transitions (1984–2000 2000–2015) simulate state 2030 validation metrics McNemar's test. introduction stochasticity into helps create typical randomness inherent real world deriving forms through iterations. high accuracy obtained experiment implies substantial fit predicted reference data. outcome proves robustness method