Bayesian Probabilistic Approach for the Correlations of Compression Index for Marine Clays

摘要: The compression index is an important soil property that essential to many geotechnical designs. Over the decades, a number of empirical correlations have been proposed relate compressibility other properties, such as liquid limit, plasticity index, in situ water content, void ratio, specific gravity, etc. reliability and thus predictability these are always being questioned. Moreover, selection between simple complicated models difficult task often depends on subjective judgments. A more model obviously provides "better fit" data but not necessarily offers acceptable degree robustness measurement noise modeling error. In present study, Bayesian probabilistic approach for class used revisit multivariate linear regression formula index. criterion structure based plausibility formulas conditional measurement, instead considering likelihood only. balances fitting capability sensitivity error, which quantified by Ockham factor. method applied analyze set 795 records, including well-known properties marine clay samples collected from various sites South Korea. It turns out correlation linking initial ratio limit possesses highest among total 18 candidate classes formulas. physical significance this most plausible addressed. be consistent with previous studies confirmation another angle.