Wildfires pose significant risks to ecosystems, human lives, and infrastructure, necessitating advanced predictive tools to mitigate their impacts. This study presents a machine learning-based framework for wildfire susceptibility mapping (WSM), designed as a predictive tool for wildfire occurrence. Using geographical information systems (GIS), a comprehensive dataset was developed by combining fourteen critical factors, including climatic, topographic, vegetation, and human activity data, from diverse sources. Four ML methods—Random Forest (RF), Support Vector Machine (SVM), Neural Network (NN), and XGBoost—were applied and compared. The results show that the XGBoost model (with an AUC of 0.96) generated the best susceptibility map. Validation using 2024–2025 fire occurrences (MODIS and Protection Civile data) showed that 87.73% of fire events were correctly captured within high and very high susceptibility zones, confirming the robustness of the proposed model. Feature importance analysis revealed that human activities, precipitation, and temperature were the most influential in wildfire prediction. These findings provide valuable insights into wildfire dynamics and contribute to the development of more effective fire prevention and mitigation strategies.

Humanitarian crises such as natural disasters and armed conflicts are increasing in frequency and intensity, posing major challenges to the sustainable protection of vulnerable populations. Rapid and equitable identification of individuals at highest risk is essential for efficient allocation of limited emergency resources and for strengthening community resilience. This study proposes an intelligent, privacy-aware decision-support platform for citizen-level vulnerability assessment that supports social sustainability and resilient crisis management. The platform integrates heterogeneous data from healthcare institutions, municipal civil records, and emergency rescue services to construct multidimensional vulnerability profiles based on social conditions, medical status, and geographical accessibility. The dataset was collected in Algeria in collaboration with the Algerian Civil Protection and consolidated into a fully anonymised dataset of approximately 5000 individual records reflecting realistic crisis scenarios. Five supervised machine learning models (Decision Tree, Random Forest, Support Vector Machine (RBF), XGBoost, and Logistic Regression) were evaluated under class-imbalance conditions using SMOTE and class weighting. The Random Forest model achieved the best performance, with an F1-Macro score of 0.710 and a recall of 0.569 for the high-risk class (95% confidence interval: [0.431, 0.706]). These results demonstrate that the proposed platform enables transparent, data-driven prioritisation of emergency interventions, contributing to sustainable humanitarian response, improved public resource allocation, and enhanced resilience of vulnerable communities.

Digital infrastructures are increasingly exposed to diverse and evolving cyber threats, highlighting the need for robust intrusion detection systems (IDSs). Although machine learning (ML)-based IDSs have achieved strong performance, most existing frameworks are still developed and evaluated mainly under intra-dataset settings, providing limited evidence of their ability to generalize across unseen environments. Moreover, few studies go beyond simply reporting cross-dataset performance to propose dedicated mechanisms for improving generalization. To address this limitation, we propose BHM-IDS, a three-stage hierarchical intrusion detection framework that combines behavior-driven hierarchy with multi-dataset training to improve generalization. The first stage performs binary detection of benign versus malicious traffic, while the second stage classifies malicious traffic into two behaviorally distinct groups: the first corresponding to flood and exhaustion attacks and the second to infiltration and exploitation attacks. The final stage performs fine-grained attack classification through two specialized multi-class classifiers. To expose the framework to more diverse attacks, CIC-IDS2017 is enriched with CIC-DDoS2019 during training, while CSE-CIC-IDS2018 is used as an external test dataset to evaluate generalization. The cross-dataset validation results yielded stage-wise accuracies of 0.93, 0.96, and 0.99, respectively, while the complete end-to-end framework achieved a weighted recall of 0.93. Recall values ranging from 0.76 to 1.00 were obtained for several major classes, including benign traffic, Patator, DoS, and DDoS, although limitations remained for certain attack categories, particularly Web Attack. Overall, the proposed framework demonstrated promising and competitive performance compared with simpler frameworks and existing state-of-the-art approaches. These findings highlight the potential of combining behavior-driven hierarchical classification with multi-dataset training to improve cross-dataset generalization in IDSs.