An interpretable model for the susceptibility of rainfall-induced shallow landslides based on SHAP and XGBoost

Geocarto International, Journal Year: 2022, Volume and Issue: 37(26), P. 13419 - 13450

Published: May 12, 2022

The machine-learning "black box" models, which lack interpretability, have limited application in landslide susceptibility mapping. To interpret the black-box some interpretable machine learning algorithms been proposed recently. Among them is SHaply Additive ExPlanation (SHAP), has attracted much attention because of its ease operation and comprehensiveness. In this study, a novel model based on SHAP XGBoost to landslides evaluation at global local levels. established provided 0.75 accuracy 0.83 AUC value for test sets. interpretation shows that peak rainfall intensity elevation are dominant factors influence occurrence study area. combination field investigations can provide comprehensive framework evaluating designated landslides, it also be used as reference preventing managing hazards landslides.

Language: Английский

---

Validity evaluation of a machine-learning model for chlorophyll a retrieval using Sentinel-2 from inland and coastal waters

Ecological Indicators, Journal Year: 2022, Volume and Issue: 137, P. 108737 - 108737

Published: March 3, 2022

The MultiSpectral Instrument (MSI) on-board Sentinel-2 provides satellite images at spatiotemporal resolutions suitable for chlorophyll a (Chla) retrieval from inland and coastal waters. Machine-learning (ML) algorithms including light gradient boosting machine (LGBM) were employed Chl MSI. study area encompasses 78 lakes estuaries located across four major river watersheds in South Korea. Matchup data between MSI overpass near-concurrent situ measurements December 2018 to April 2021 included. remote sensing reflectance (Rrs) values of six single spectral bands two-band ratios used as the input features. Despite difficulty Chla estimation optically complex waters, ML showed overall reasonable accuracy. Among algorithms, LGBM exhibited best performance (R2 = 0.75, bias -0.15, slope 0.73, RMSE 15.15 mg·m-3, MAE 9.49 mg·m-3) over wide range trophic states. Post-hoc interpretations performing using Shapley additive explanations indicated that Rrs(7 0 4)/Rrs(6 6 5) was most important feature, while 3 9)/Rrs(7 4) Rrs(4 9 2)/Rrs(5 0) played auxiliary roles through interaction with 5). Among-lake spatial variations explained by percent forest agricultural within buffer zone multiple scales (buffer widths 50 m 500 m). associations modeled land cover types, is, increase concentration decrease area, consistent established ecological knowledge. Overall, model among confirmed validity retrieving MSI-derived estuaries. Our can serve reference evaluating models water sensing.

Language: Английский

---

Interpretable machine learning with tree-based shapley additive explanations: Application to metabolomics datasets for binary classification

PLoS ONE, Journal Year: 2023, Volume and Issue: 18(5), P. e0284315 - e0284315

Published: May 4, 2023

Machine learning (ML) models are used in clinical metabolomics studies most notably for biomarker discoveries, to identify metabolites that discriminate between a case and control group. To improve understanding of the underlying biomedical problem bolster confidence these model interpretability is germane. In metabolomics, partial least square discriminant analysis (PLS-DA) its variants widely used, partly due model's with Variable Influence Projection (VIP) scores, global interpretable method. Herein, Tree-based Shapley Additive explanations (SHAP), an ML method grounded game theory, was explain local explanation properties. this study, experiments (binary classification) were conducted three published datasets using PLS-DA, random forests, gradient boosting, extreme boosting (XGBoost). Using one datasets, PLS-DA explained VIP while best-performing models, forest model, interpreted Tree SHAP. The results show SHAP has more depth than PLS-DA's VIP, making it powerful rationalizing machine predictions from studies.

Language: Английский

---

Assessment of Wildfire Susceptibility and Wildfire Threats to Ecological Environment and Urban Development Based on GIS and Multi-Source Data: A Case Study of Guilin, China

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(10), P. 2659 - 2659

Published: May 19, 2023

The frequent occurrence and spread of wildfires pose a serious threat to the ecological environment urban development. Therefore, assessing regional wildfire susceptibility is crucial for early prevention formulation disaster management decisions. However, current research on primarily focuses improving accuracy models, while lacking in-depth study causes mechanisms wildfires, as well impact losses they cause This situation not only increases uncertainty model predictions but also greatly reduces specificity practical significance models. We propose comprehensive evaluation framework analyze spatial distribution effects influencing factors, risks damage local In this study, we used information from period 2013–2022 data 17 factors in city Guilin basis, utilized eight machine learning algorithms, namely logistic regression (LR), artificial neural network (ANN), K-nearest neighbor (KNN), support vector (SVR), random forest (RF), gradient boosting decision tree (GBDT), light (LGBM), eXtreme (XGBoost), assess susceptibility. By evaluating multiple indicators, obtained optimal Shapley Additive Explanations (SHAP) method explain decision-making mechanism model. addition, collected calculated corresponding with Remote Sensing Ecological Index (RSEI) representing vulnerability Night-Time Lights (NTLI) development vulnerability. coupling results two represent ecology city. Finally, by integrating information, assessed risk disasters reveal overall characteristics Guilin. show that AUC values models range 0.809 0.927, ranging 0.735 0.863 RMSE 0.327 0.423. Taking into account all performance XGBoost provides best results, AUC, accuracy, 0.863, 0.327, respectively. indicates has predictive performance. high-susceptibility areas are located central, northeast, south, southwest regions area. temperature, soil type, land use, distance roads, slope have most significant Based assessments, potential can be identified comprehensively reasonably. article improve prediction provide important reference response wildfires.

Language: Английский

---

Optimisation and interpretation of machine and deep learning models for improved water quality management in Lake Loktak

Journal of Environmental Management, Journal Year: 2023, Volume and Issue: 351, P. 119866 - 119866

Published: Dec. 25, 2023

Language: Английский

---

Machine learning and explainable AI for chlorophyll-a prediction in Namhan River Watershed, South Korea

Ecological Indicators, Journal Year: 2024, Volume and Issue: 166, P. 112361 - 112361

Published: July 16, 2024

Algal blooms are a primary concern in freshwater quality management. Thus, prediction of algal concentrations is crucial. Chlorophyll-a (Chl-a) an indicator concentration. This study focuses on the downstream watershed Namhan River, which significant water source for Korean metropolitan area. Using 25 input variables, we developed eXtreme Gradient Boosting (XGB) model predicting Chl-a Yanpyeong. The XGB exhibited impressive predictability (R2 = 0.9487, RMSE 3.1661, RSR 0.2781). To assess variations based tree-model-based Feature Importance (Tree-FI) and Shapley Additive exPlanation (SHAP)-based feature importance (SHAP-FI) were used. validates utility eXplainable Artificial Intelligence (XAI) through SHAP Partial Dependency Plot (PDP) analyses, revealing positive contributions pH turbidity Yangpyeong, Hongcheon, to concentrations. Additionally, it identifies complex interactions between variables affecting concentrations, emphasizing intricate relationship bloom research underscores significance integrating machine learning models XAI techniques addressing real-world environmental challenges, providing valuable tools effective prevention management strategies.

Language: Английский

---

AI and machine learning in climate change research: A review of predictive models and environmental impact

World Journal of Advanced Research and Reviews, Journal Year: 2024, Volume and Issue: 21(1), P. 1999 - 2008

Published: Jan. 25, 2024

The burgeoning threat of climate change has spurred an increased reliance on advanced technologies to comprehend and mitigate its far-reaching consequences. Artificial Intelligence (AI) Machine Learning (ML) have emerged as indispensable tools in research, offering unprecedented capabilities for predictive modeling assessing environmental impact. This review synthesizes the current state AI ML applications emphasizing their role understanding repercussions. Predictive models leveraging algorithms demonstrated remarkable efficacy forecasting patterns, extreme weather events, sea-level rise. These incorporate vast datasets encompassing meteorological, geospatial, oceanic information, enabling more accurate predictions future scenarios. Moreover, AI-driven excel recognizing intricate patterns non-linear relationships within data, enhancing capacity simulate complex systems. Environmental impact assessment stands a critical facet techniques are proving instrumental this regard. facilitate analysis diverse ecological parameters, including deforestation rates, biodiversity loss, carbon sequestration dynamics. By discerning nuanced immense datasets, systems contribute direct indirect consequences ecosystems. Despite these advancements, challenges persist, such need standardized data formats, model interpretability, ethical considerations. Additionally, integration findings into policy frameworks remains crucial frontier. As intersection AI, ML, research evolves, continuous interdisciplinary collaboration is essential harness full potential safeguarding our planet's future. illuminates landscape applications, providing insights efficacy, challenges, contributions advancing sustainability.

Language: Английский

---

A SHAP-Enhanced XGBoost Model for Interpretable Prediction of Coseismic Landslides

Advances in Space Research, Journal Year: 2024, Volume and Issue: 74(8), P. 3826 - 3854

Published: July 9, 2024

Language: Английский

---

Unveiling the Hidden Connections: Using Explainable Artificial Intelligence to Assess Water Quality Criteria in Nine Giant Rivers

Journal of Cleaner Production, Journal Year: 2025, Volume and Issue: unknown, P. 144861 - 144861

Published: Jan. 1, 2025

Language: Английский

---

Data-driven models for predicting community changes in freshwater ecosystems: A review

Ecological Informatics, Journal Year: 2023, Volume and Issue: 77, P. 102163 - 102163

Published: June 15, 2023

Language: Английский

---