Brownfield redevelopment and land sustainable management have many interconnected aspects, which could be evaluated by regional development. Most contaminated sites in Taiwan often lack comprehensive consideration during remediation for land reuse planning; therefore, they are operated independently from regional development. Regional development is a dynamic process; however, there is a research gap in understanding how regional development can serve as an incentive for brownfield redevelopment. This study established a model for assessing regional sustainability using machine learning (ML-RSD), that model used principal component analysis (PCA) and support vector machine (SVM) to predict the comprehensive evaluation rating of regional development at the township level and to explore the trends of regional governance sustainability. The results of ML-RSD that are compared to commonly used methods (driver-pressure-state-impact-response and entropy weight method) were consistent. Indicators such as the protection of vulnerable groups (SDG1), food security (SDG2), good health and well-being (SDG3), inclusive and equitable quality education (SDG4), environmental protection (SDG6), accessible and reliable energy (SDG7), and sustainable economic growth (SDG8) were identified as the main factors affecting regional development. Additionally, the model achieved an accuracy value exceeding 80%. Additionally, the results showed that there was no significant correlation between the delisted site (completed remediation) and regional development in southern Taiwan. The model underscored the overall lower comprehensive sustainable values for regions affected by inadequate land rehabilitation governance (SDG15). The results present that brownfield redevelopment strategies should enhance their effectiveness in land adaptation and integrating brownfield redevelopment with regional sustainable development is essential for it to become a win–win solution.
Read full abstract