Portal de Periódicos da CAPES

An Artificial Intelligence Approach Based on Multi-layer Perceptron Neural Network and Random Forest for Predicting Maximum Dry Density and Optimum Moisture Content of Soil Material in Quang Ninh Province, Vietnam

2021; Springer Nature; Linguagem: Inglês

10.1007/978-981-16-7160-9_176

2366-2557

Manh Nguyen Duc, An Ho Sy, Truong Nguyen Ngoc, Thuy Linh Hoang Thi,

Landslides and related hazards

Maximum dry density (ρd(max)) and optimum moisture content (wopt) are two key parameters of embankment fill soil material using in transport construction. To obtain these parameters, Proctor test (ASTM D698/AASHTO) or Modified Proctor test (ASTM D1557/AASHTO T180 etc.) is traditionally performed in the laboratory. However, this test takes time and expenses. Moreover, the accuracy of the test depends significantly on the collection of samples, expertize of the testers and quality of the experimental apparatuses. In this study, the main aim is to propose two machine learning approaches named Multi-layer Perceptron Neural Network (ANN-MLP) and Random Forest (RF) for the prediction of ρd(max) and wopt. Input parameters include silt content(%), clay content (%), liquid limit (%), plastic limit (%), plasticity index (%), specific gravity which have strong correlations with ρd(max) and wopt were used in the model. Performance of the model was assessed by statistical methods, such as Mean Absolute Error (MAE), Root mean square error (RMSE), and Coefficient of determination (R2). Results of the models study indicate that the proposed models ANN-MLP and RF has the same predictive capability (R2average of ANN-MLP is 0.829 and R2average of RF is 0.827). The results of this study might help in quickly predicting ρd(max) and wopt of embankment fill soil material.