Assessment of earthquake-induced liquefaction susceptibility using ensemble learning

Madhav, Madhira (2023) Assessment of earthquake-induced liquefaction susceptibility using ensemble learning. Multiscale and Multidisciplinary Modeling, Experiments and Design, 6 (2). pp. 251-261. ISSN 2520-8179

Full text not available from this repository.

Abstract

Liquefaction is one of the most hazardous phenomena which occurs because of the earthquake and causes great loss of constructed facilities and mortality. Therefore, it is essential to consider the liquefaction susceptibility of the site at an earlier stage in any construction project. In this study, ensemble learning is used for predicting liquefaction susceptibility. Pearson’s correlation analysis is adopted for the selection of the input parameters using a dataset consisting of 620 data. Various artificial neural networks and random forest classifier models are trained and validated using nine geotechnical and seismic parameters. The performance of developed models in this study is examined using three statistical parameters, i.e., accuracy, confusion matrix, and precision. The random forest classifier model is found to be suitable for the prediction of the liquefaction susceptibility with an accuracy of 98.38%. The permutation analysis is performed to find out the relative importance of the input parameters and it is found that liquefaction susceptibility is highly susceptible to SPTN and F ≤ 75 µm.

[error in script]
IITH Creators:
IITH CreatorsORCiD
Item Type: Article
Uncontrolled Keywords: Artificial neural network; Earthquake; Geotechnical engineering; Liquefaction; Random forest classifier; Earthquakes; Classifier models; Construction projects; Correlation analysis; Ensemble learning; Geotechnical parameters; Input parameter; Liquefaction susceptibility; Performance; Random forest classifier; Seismic parameters; Forecasting; Forestry; Geotechnical engineering; Neural networks; Soil liquefaction
Subjects: Physics
Physics > Fluid mechanics
Divisions: Department of Physics
Depositing User: Mr Nigam Prasad Bisoyi
Date Deposited: 06 Jan 2024 12:20
Last Modified: 06 Jan 2024 12:20
URI: http://raiith.iith.ac.in/id/eprint/11784
Publisher URL: https://doi.org/10.1007/s41939-023-00146-z
OA policy: https://v2.sherpa.ac.uk/id/publication/36143
Related URLs:

    Actions (login required)

    View Item View Item
    Statistics for RAIITH ePrint 11784 Statistics for this ePrint Item