Modeling Susceptibility to Illegal Land Occupation in Cape Town, South Africa: A Random Forest Approach

ABSTRACT Illegal land occupation is a pervasive challenge in South Africa, particularly in urban landscapes such as Cape Town. Despite numerous studies investigating its causes, spatial distribution, and impacts, there is a scarcity of research focused on predicting its occurrence. This study addresses this gap by developing a machine learning model to identify areas susceptible to illegal land invasion within the Cape Town Metropole. A random forest (RF) regression model was built utilizing samples of prevalence and pseudoabsence. A total of 1070 samples were collected (535 points per class), with 70% used for training and 30% used for validation. The RF model demonstrated accurate predictive performance with an R 2 of 0.809. The study's findings indicate that areas bordering existing informal settlements are most at risk of future illegal land occupation. The developed RF model can be a valuable tool for urban planners and policymakers in developing more effective strategies to prevent and manage illegal land occupation.