Hybrid Model: Teaching Learning-Based Optimization of Artificial Neural Network (TLBO-ANN) for the Prediction of Soil Permeability Coefficient

The permeability coefficient (k-value) of the soil is an important parameter used in civil engineering design roads, tunnels, dams, and other structures. However, determination k-value by experimental methods laboratory or field still costly time-consuming. Moreover, it requires special equipment care collection samples for study. Therefore, this study, we have proposed machine learning (ML) hybrid model: teaching learning-based optimization artificial neural network (TLBO-ANN) to predict based on limited parameters (natural water content, void ratio, specific gravity, liquid limit, plastic clay content) which can be determined easily laboratory. Test results 84 obtained from Da Nang-Quang Ngai expressway project Vietnam are model development. Statistical indicators such as correlation (R), root mean square error (RMSE), absolute (MAE) validate evaluate accuracy model. show that TLBO-ANN effective tool predicting correctly (R = 0.905) consideration structures founded soil.