The cohesion and internal friction angle of loess are important macro-mechanical parameters for evaluating the safety and stability of engineering construction. Traditional laboratory measurement methods are time-consuming and difficult to conduct on-site. This study aims to compare the effectiveness of five Machine Learning (ML) methods, namely Random Forest (RF), Support Vector Machine (SVM), Back Propagation Neural Network (BPNN), BPNN optimized by Particle Swarm Optimization (PSO-BPNN) and BPNN optimized by Genetic Algorithm (GA-BPNN), in predicting the macro-mechanical properties of loess. To this end, the study collected data from 89 undisturbed loess samples and 229 remolded loess samples to construct training and testing datasets, and used three correlation analysis methods to analyze the influence of physical parameters on mechanical properties. The study found that the water content has the most significant impact on the mechanical properties of loess. In terms of prediction ability, SVM performs the best among the ML methods used, and the determination coefficient for cohesion of undisturbed loess reaches 0.857. Although the training data is limited, the prediction performance of BPNN is significantly improved after being optimized by PSO or GA. The research results show that ML provides an effective way to study the complex mechanical behavior of loess.

The rapid progress of urbanization and industrialization has led to the accumulation of large amounts of metal ions in the environment. These metal ions are adsorbed onto the negatively charged surfaces of clay particles, altering the total surface charge, double-layer thickness, and chemical bonds between the particles, which in turn affects the interactions between them. This causes changes in the microstructure, such as particle rearrangement and pore morphology adjustments, ultimately altering the mechanical behavior of the soil and reducing its stability. This study explores the effects of four common metal ions, including monovalent alkali metal ions (Na+, K+) and divalent heavy metal ions (Pb2+, Zn2+), with a focus on how ion valence and concentration impact the soil's microstructure and mechanical properties. Microstructural tests show that metal ion incorporation reduces particle size, increases clay content, and transforms the structure from layered to honeycomb-like. Small pores decrease while large pores dominate, reducing the specific surface area and pore volume, while the average pore size increases. Although cation exchange capacity decreases, cation adsorption density per unit surface area increases. Monovalent ions primarily disperse the soil structure, while divalent ions induce coagulation. Macro-mechanical tests reveal that metal ion contamination reduces porosity under loading, with compressibility rises as the ion concentration increases. Soils contaminated with alkali metal ions shows higher compression coefficients at all loads, while heavy metal ions cause higher compression under lower loads. Shear strength, the internal friction angle, and cohesion in metal-ion-contaminated clay decrease compared to uncontaminated field-state clay, with greater declines at higher ion concentrations. The Micropore Morphology Index and hydro-pore structural parameter effectively characterize both micro- and macrostructural properties, establishing a quantitative relationship between HPSP and the engineering properties of metal-ion-contaminated clay.