Strength Prediction by Support Vector Regression (SVR) for Biopolymer-Based Soil Treatment (BPST)

Abstract

For sustainable and environmentally friendly geotechnical engineering, biological soil treatment methods have recently been actively introduced. Biopolymer-based soil treatment (BPST) is regarded as a low carbon footprint ground improvement method with adequate strengthening and pore clogging characteristics. However, no detailed analysis of the relationship between the components that have a significant impact on BPST strengthening behavior has been carried out. Because BPST is frequently implemented as a mixture of soil, biopolymer, and water, different results can be obtained depending on the hydrogel phase and soil type. Using unconfined compressive strength (UCS) data for BPST obtained from laboratory test, support vector regression (SVR) was used in this study to predict the UCS and examine the dominant characteristics influencing the strengthening behavior of biopolymer-treated soil mixes.