Predicting deep-seated landslide displacement on Taiwan's Lushan through the integration of convolutional neural networks and the Age of Exploration-Inspired Optimizer

A groundbreaking study achieves high-precision prediction of deep-seated landslide displacements in Taiwan's Lushan National Park, with a mean absolute percentage error (MAPE) of just 2.81%. The research integrates convolutional neural networks (CNNs) and the Age of Exploration-Inspired Optimizer (AEIO), leveraging eight years of displacement, groundwater, and meteorological data to enhance geohazard risk assessment. This innovative framework, combining machine learning and time series deep learning, offers a robust tool for early warning systems, significantly improving the safety of road networks, construction projects, and infrastructure in vulnerable slope areas. The AEIO-MobileNet model's accuracy marks a substantial advancement in landslide prediction, addressing critical needs for environmental risk management amid changing climate patterns.

Chou, J.-S., Nguyen, H.-M., Phan, H.-P., & Wang, K.-L. (2025). Predicting deep-seated landslide displacement on Taiwan’s Lushan through the integration of convolutional neural networks and the Age of Exploration-Inspired Optimizer. Natural Hazards and Earth System Sciences, 25(1), 119–146. https://doi.org/10.5194/nhess-25-119-2025