Landslides remain one of the most critical natural hazards, posing significant threats to infrastructure, the environment, and human life. Traditional approaches to landslide risk prediction, such as rainfall threshold models and image-based classification, often face limitations including data imbalance, low generalizability, and poor performance in capturing medium-to high-risk scenarios. This study introduces a predictive framework that integrates synthetic data generation with a multiple logistic regression model to improve landslide risk assessment in the Malaysian context. The model was trained on balanced datasets and evaluated through confusion matrices, performance metrics, and validation using unseen data across three distinct scenarios. Results demonstrate that a multiple-logistic-regression model trained on this balanced data achieved an overall accuracy of 0.73, precision of 0.73, recall of 0.73, and a Receiver Operating Characteristic-Area Under Curve (ROC-AUC) of 0.80. In three validation scenarios using unseen data from 2015-2024 (three months before, during, and three months after known landslide events), the model correctly identified medium and high-risk periods when other machine-learning models defaulted to low-risk predictions. The study highlights the trade-off between accuracy and generalization in machine-learning-based early warning systems and underscores the importance of class-balancing and rigorous validation for real-world applicability. Our findings, therefore, demonstrate that the logistic-regression model, when paired with synthetic data augmentation, can serve as a cost-effective, interpretable pre-screening tool for regional landslide risk assessment in Malaysia.
This paper aims to examine the association of the specific governance indicators, which are the regulatory quality, the rule of law, and government effectiveness, on sustainable development in 60 countries around the world in 2024. This is explained by the key role that the state institutions and the institutional framework may play in enhancing economic, social, and environmental results and in achieving the Sustainable Development Goals (SDGs). The study employed the quantitative approach that is based on 2024 cross-sectional data. The data were obtained from the 2024 SDG Index and the World Bank's Worldwide Governance Indicators (WGI). The Eviews software was used to compute an Ordinary Least Squares (OLS) multiple linear regression model to examine the relationship between the independent variables (regulatory quality, the rule of law and government effectiveness) and the dependent variable (the SDG Index). The results reflected that the rule of law and the efficacy of the government have a positive and substantial effect on sustainable development, but the regulatory quality did not show a direct significant impact. This shows that sustainable development is based on the unity of the institutional framework which consolidates legal, regulatory, and administrative potential to achieve quantitative results.
