On Random Subspace Optimization-Based Hybrid Computing Models Predicting the California Bearing Ratio of Soils

The California Bearing Ratio (CBR) is an important index for evaluating the bearing capacity of pavement subgrade materials. In this research, random subspace optimization-based hybrid computing models were trained and developed for the prediction of the CBR of soil. Three models were developed, namely reduced error pruning trees (REPTs), random subsurface-based REPT (RSS-REPT), and RSS-based extra tree (RSS-ET). An experimental database was compiled from a total of 214 soil samples, which were classified according to AASHTO M 145, and included 26 samples of A-2-6 (clayey gravel and sand soil), 3 samples of A-4 (silty soil), 89 samples of A-6 (clayey soil), and 96 samples of A-7-6 (clayey soil). All CBR tests were performed in soaked conditions. The input parameters of the models included the particle size distribution, gravel content (G), coarse sand content (CS), fine sand content (FS), silt clay content (SC), organic content (O), liquid limit (LL), plastic limit (PL), plasticity index (PI), optimum moisture content (OMC), and maximum dry density (MDD). The accuracy of the developed models was assessed using numerous performance indexes, such as the coefficient of determination, relative error, MAE, and RMSE. The results show that the highest prediction accuracy was obtained using the RSS-based extra tree optimization technique.

Group velocity maps using subspace and Trans-dimensional inversions: ambient noise tomography in the Western part of Java, Indonesia

Summary In this paper, we compare two different methods for group velocity inversion: iterative, least-squares subspace optimization, and probabilistic sampling based on the Trans-dimensional Bayesian method with tree-based wavelet parameterization. The wavelet parameterization used a hierarchical prior for wavelet coefficients which could adapt to the data. We applied these inversion methods for ambient noise tomography of the western part of Java, Indonesia. This area is an area prone to multiple geological hazards due to its proximity to the subduction of the Australia Plate beneath Eurasia. It is therefore important to have a better understanding of upper crustal structure to support seismic hazard and disaster mitigation efforts in this area. We utilized a new waveform dataset collected from 85 temporary seismometers deployed during 2016–2018. Cross-correlation of the waveform data was applied to retrieve empirical Rayleigh wave Green's functions between station pairs, and the spatial distribution of group velocity was obtained by inverting dispersion curves. Our results show that, although computationally expensive, the Trans-dimensional Bayesian approach offered important advantages over optimization, including more effective explorative of the model space and more robust characterization of the spatial pattern of Rayleigh wave group velocity. Meanwhile, the iterative, least-square subspace optimization suffered from the subjectivity of choice for reference velocity model and regularization parameter values. Our Rayleigh wave group velocity results show that for short (1–10 s) periods group velocity correlates well with surface geology, and for longer periods (13–25 s) it correlates with centers of volcanic activity.