Machine learning-based techniques for land subsidence simulation in an urban area

Liu, Jianxin; Liu, Wenxiang; Allechy, Fabrice Blanchard; Zheng, Zhiwen; Liu, Rong; Kouadio, Kouao Laurent

doi:10.1016/j.jenvman.2024.120078

Cited by 6 publications

(4 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To prepare the input variables, Sentinel-1 satellite images of the study areas for 2015-2017 were obtained, and the locations and dimensions of several subsidence points in the areas were extracted using image analysis and by referring to technical reports (obtained from the Geological Survey and Mineral Exploration Institute) and previous studies [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31]. The geographic location of the occurrence and the non-occurrence LS points (provided by the Geological Survey and Mineral Exploration Institute, the abovementioned studies, and RS-GIS-based methods) are used to obtain LS maps.…”

Section: Data Gatheringmentioning

confidence: 99%

“…They reported that MARS outperformed other MLAs in the study area. Liu et al [20] addressed LS in urban planning and infrastructure management by using two machine learning models, including the extreme gradient boosting regressor (XGBR) and long short-term memory (LSTM). They identified groundwater level (GWL) and building concentration (BC) as key factors influencing LS.…”

Section: Introductionmentioning

confidence: 99%

“…To achieve these aims, we scrutinized the performance of six classification MLAs (classification and regression tree (CART), Bayesian linear regression (BLR), SVM, boosted BRT, RF, and logistic regression (LogR)) and one regression method (multiple linear regression (MLR)) in identifying LS-prone areas and predicting the magnitude of LS. We applied the methods to two study areas, Semnan Plain and Kashmar Plain in Iran, both of which have experienced severe LS in recent decades [16][17][18][19][20][21][22][23][24], using nine input variables: distance from the river, distance from the fault, groundwater drawdown, slope, aspect, land use, lithology, topographic wetness index (TWI), and normalized difference vegetation index (NDVI).…”

Section: Introductionmentioning

confidence: 99%

“…These seven MLAs were selected for analysis based on their performance merits and previous studies [16][17][18][19][20][21][22][23][24][25][26][27]. The CART approach can handle missing values in the training dataset and is insensitive to outliers since it uses surrogates [28].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evaluating Machine Learning-Based Approaches in Land Subsidence Susceptibility Mapping

Hosseinzadeh,

Anamaghi,

Behboudian

et al. 2024

Land

View full text Add to dashboard Cite

Land subsidence (LS) due to natural and human-driven forces (e.g., earthquakes and overexploitation of groundwater) has detrimental and irreversible impacts on the environmental, economic, and social aspects of human life. Thus, LS hazard mapping, monitoring, and prediction are important for scientists and decision-makers. This study evaluated the performance of seven machine learning approaches (MLAs), comprising six classification approaches and one regression approach, namely (1) classification and regression trees (CARTs), (2) boosted regression tree (BRT), (3) Bayesian linear regression (BLR), (4) support vector machine (SVM), (5) random forest (RF), (6) logistic regression (LogR), and (7) multiple linear regression (MLR), in generating LS susceptibility maps and predicting LS in two case studies (Semnan Plain and Kashmar Plain in Iran) with varying intrinsic characteristics and available data points. Multiple input variables (slope, aspect, groundwater drawdown, distance from the river, distance from the fault, lithology, land use, topographic wetness index (TWI), and normalized difference vegetation index (NDVI)), were used as predictors. BRT outperformed the other classification approaches in both case studies, with accuracy rates of 75% and 74% for Semnan and Kashmar plains, respectively. The MLR approach yielded a Mean Square Error (MSE) of 0.25 for Semnan plain and 0.32 for Kashmar plain. According to the BRT approach, the variables playing the most significant role in LS in Semnan Plain were groundwater drawdown (20.31%), distance from the river (17.11%), land use (14.98%), NDVI (12.75%), and lithology (11.93%). Moreover, the three most important factors in LS in Kashmar Plain were groundwater drawdown (35.31%), distance from the river (23.1%), and land use (12.98%). The results suggest that the BRT method is not significantly affected by data set size, but increasing the number of training set data points in MLR results in a decreased error rate.

show abstract