Comparison of Land Cover Change Prediction Models: A Case Study in Kedungkandang District, Malang City

Abstract

The infrastructure of Malang City is currently being directed towards the eastern and southeastern parts, Kedungkandang District. Infrastructure plays an important role in the aspect of land cover change, which raises the complexity of the emergence of urban forms and dynamics. This study compares three models, Artificial Neural Network (ANN), Logistic Regression (LR), and Multi-Criteria Evaluation (MCE), to predict changes in land cover in the Kedungkandang District using the Cellular Automata (CA) approach. The prediction results indicate that the ANN and MCE models have the highest overall Kappa values (prediction accuracy), while the ANN and LR models have the highest location-specific Kappa values. However, overall, the ANN model demonstrates the highest accuracy and performance among the other two models. This research makes a significant contribution to urban planning by highlighting the importance of using machine learning-based technology to predict land cover changes in Malang City, particularly in the Kedungkandang District. Stakeholders can leverage this technology to design more effective and sustainable infrastructure policies and implement preventive measures to mitigate the negative impacts of uncontrolled urban growth.