Machine Learning-Based Prediction of Unconfined Compressive Strength of Sands Treated by Microbially-Induced Calcite Precipitation (MICP): A Gradient Boosting Approach and Correlation Analysis

The current study applies a soft-computing approach based on the gradient boosting method to predict unconfined compressive strength (UCS) of sands treated with microbially-induced calcite precipitation (MICP). A 10-fold cross-validation and hyperparameter tuning are performed find optimal architecture algorithm. total 402 data compression tests biocemented utilized in this study. dataset includes eight input parameters: median sand particle size, uniformity coefficient sand, initial void ratio, calcium chloride concentration, urea urease activity, optical density bacteria, content. finding demonstrates that outperformed five commonly used machine learning algorithms (artificial neural networks, random forests, k-nearest neighbors, support vector regression, decision trees) predicting UCS sands. Using boosting, predicted has strong correlation actual values (R2 = 0.95). Moreover, series feature importance analyses carried out over dataset. relationships between strength, content, ratio discussed within article. Furthermore, some guidelines provided for assessing effect environmental factors For further study, limitations regarding insufficiency modification addressed.