LSTM-Based Prediction of Active Earth Pressure Coefficient Against Rigid Retaining Walls

The application of a Long Short-Term Memory (LSTM) neural networks have been explored in this study to predict the active earth pressure coefficient (AEPC). The AEPC is a critical parameter in geotechnical engineering for designing retaining structures. Traditional methods rely on empirical formulas and manual calculations, which may lack accuracy and efficiency. By leveraging the capabilities of LSTM neural networks, which are well-suited for sequential data analysis and long-term dependencies, this research aims to develop a more accurate and reliable prediction model. The performance of the LSTM model is evaluated based on metrics such as Mean Squared Error (MSE) and determination coefficient (R 2 ), and score function (S) to ensure its predictive accuracy. The obtained experimental results (MSE = 0.0003, R 2 = 0.97, and S = 0.0014), even for a limited number for training epochs, demonstrate that the LSTM neural networks can effectively outperforms and predict the active earth pressure coefficient with high precision compared to conventional methods widely used in this field.