Application of time series prediction techniques for coastal bridge engineering

Abstract In this study, three machine learning techniques, the XGBoost (Extreme Gradient Boosting), LSTM (Long Short-Term Memory Networks), and ARIMA (Autoregressive Integrated Moving Average Model), are utilized to deal with time series prediction tasks for coastal bridge engineering. The performance of these techniques is comparatively demonstrated in typical cases, wave-load-on-deck under regular waves, structural displacement combined wind wave loads, height variation along typhoon/hurricane approaching. To enhance accuracy, a data preprocessing method adopted an improved framework model after rolling forecast proposed. obtained results show that: (a) When making on featured periodic regularity, both models perform well, can make predictions multi-step ahead, (b) predict just one step ahead based aperiodic dataset limited amplitude more accurately, while appropriate preprocessing, (c) All tendency updating over time, but accuracy favorable. successful application provide guidance resolve engineering problems time-history requirements.