Research on Bayesian network probabilistic assessment model for soil seismic liquefaction considering small-sample and incomplete SPT data

Liquefaction assessment is a critical task in seismic geotechnical engineering; however, existing methods are often limited by their reliance on complete datasets and inability to effectively handle missing information. This study proposes a Bayesian network-based framework for probabilistic liquefaction assessment under conditions of small sample size and incomplete Standard Penetration Test ( SPT ) data. Causal logical relationships derived from liquefaction mechanisms are incorporated to construct interpretable network structures. A systematic experimental framework integrating different parameter learning algorithms, training strategies, and missing-data handling schemes is developed to evaluate model performance under varying conditions. The results illustrate that small-sample models show limited performance, whereas incorporating incomplete information significantly improves prediction accuracy. Under the optimal combination of the real-sample-based training strategy (Method-1) and the maximum a posteriori probability (MAP)-based missing-data handling scheme (Scheme II), the five-factor Bayesian network model achieved an Overall Accuracy ( OA ), Precision ( Pre ), Recall ( Rec ), and F 1 -score ( F 1 ) of 80.56%, 74.00%, 97.37%, and 84.09%, respectively, on the independent validation dataset. It demonstrates the reliable predictive capability of the proposed model for unseen liquefaction cases. The proposed framework expands the applicability of probabilistic liquefaction assessment across different depth ranges, seismic loading conditions, and soil properties. Moreover, it provides an interpretable and robust solution for engineering scenarios where data incompleteness is unavoidable, offering practical value for seismic hazard assessment.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-17
DOI
https://doi.org/10.1038/s41598-026-80717-5
Primary Topic
Geotechnical Engineering and Soil Mechanics
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Research on Bayesian network probabilistic assessment model for soil seismic liquefaction considering small-sample and incomplete SPT data

Meng Fan, Zhengquan Yang, Jingjun Li, Jianming Zhao et al.
2 links
Scientific Reports
Geotechnical Engineering and Soil Mechanics
article

Research on Bayesian network probabilistic assessment model for soil seismic liquefaction considering small-sample and incomplete SPT data

Meng Fan, Zhengquan Yang, Jingjun Li, Jianming Zhao, Xiaosheng Liu
article en

Abstract

Liquefaction assessment is a critical task in seismic geotechnical engineering; however, existing methods are often limited by their reliance on complete datasets and inability to effectively handle missing information. This study proposes a Bayesian network-based framework for probabilistic liquefaction assessment under conditions of small sample size and incomplete Standard Penetration Test ( SPT ) data. Causal logical relationships derived from liquefaction mechanisms are incorporated to construct interpretable network structures. A systematic experimental framework integrating different parameter learning algorithms, training strategies, and missing-data handling schemes is developed to evaluate model performance under varying conditions. The results illustrate that small-sample models show limited performance, whereas incorporating incomplete information significantly improves prediction accuracy. Under the optimal combination of the real-sample-based training strategy (Method-1) and the maximum a posteriori probability (MAP)-based missing-data handling scheme (Scheme II), the five-factor Bayesian network model achieved an Overall Accuracy ( OA ), Precision ( Pre ), Recall ( Rec ), and F 1 -score ( F 1 ) of 80.56%, 74.00%, 97.37%, and 84.09%, respectively, on the independent validation dataset. It demonstrates the reliable predictive capability of the proposed model for unseen liquefaction cases. The proposed framework expands the applicability of probabilistic liquefaction assessment across different depth ranges, seismic loading conditions, and soil properties. Moreover, it provides an interpretable and robust solution for engineering scenarios where data incompleteness is unavoidable, offering practical value for seismic hazard assessment.

Scientific Reports
China Institute of Water Resources and Hydropower Research (CN), Ministry of Water Resources of the People's Republic of China (CN)
State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin
Geotechnical Engineering and Soil Mechanics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon