Shear Strength Prediction for Fiber-Reinforced Soils by Data Mining Techniques and Their Ensembles
Other Title
應用資料探勘技術暨啟發式演化組合模型預測離散纖維加勁土壤之剪力強度參數
Journal
Journal of the Chinese Institute of Civil and Hydraulic Engineering
Journal Volume
28
Journal Issue
3
Pages
205-218
Date Issued
2016
Author(s)
Abstract
The accuracy of theoretical and empirical models for predicting the shear strength of fiber-reinforced soils (FRS) is questionable because of the difficulty of using these simplified models to describe the complex mechanism of soilfiber interaction. This study compiled a large database of available high quality triaxial and direct shear tests on FRS documented in the literature from 1983 to 2015. The database includes information on the properties of sand, fibers, soilfiber interface, and stress parameters. Data mining technologies were employed to identify factors influencing shear strength and to predict the peak friction angle of FRS. The analysis techniques included (1) classification and regression methods, e.g., linear REGression (REG) analysis, Classification And Regression Tree (CART) analysis, a GENeralized LINear (GENLIN) Model, and CHi-squared Automatic Interaction Detection (CHAID); (2) machine learners, e.g., Artificial Neural Network (ANN) and Support Vector Machine/Regression (SVM/SVR); and (3) meta ensemble models, e.g., Voting, Bagging, Stacking and Tiering. The analytical results indicated that fiber content, fiber aspect ratio, soil friction angle and stress parameter had the largest effects on FRS shear strength. The optimal model obtained after further model training, crossvalidation, and testing was the Tiering SVM-(SVR/SVR) method. The correlation coefficient of the prediction values with the measured values in the database was 0.89. The mean absolute percentage error was <4%, root mean square error was <2°, and mean absolute error was <2°. The overall improvement in performance measures was 13.12% ~ 79.50% with respect to theoretical or empirical models.
Type
journal article