Landslide Susceptibility Prediction based on Decision Tree and Feature Selection Methods

Nirbhav,; Malik, Anand; Maheshwar, .; Jan, Tony; Prasad, Mukesh

doi:10.1007/s12524-022-01645-1

Cited by 11 publications

(7 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both mappings yield worse results for both assessments using the validation and StorMe dataset in comparison to the reference map. Therefore, this result rather supports the observation by Malik et al (2023). Taking into account the just described observations in the present paper and the typically limited amount of features chosen in landslide susceptibility mapping studies way below the number of added irrelevant features in Kuhn and Johnson (2019), supports the claim that the focus should be on the integration of balanced and physically meaningful features into the mapping process instead of focusing on reducing the training dataset as much as possible or the conduction of intensive hyperparameter tuning.…”

Section: Discussionsupporting

confidence: 87%

“…Pham et al (2021) also describe the desire to perform feature selection to reduce computing time and increase accuracy of the ML model. The in Malik et al (2023) observed reduction of the accuracy contrasts the result of e.g. Kuhn and Johnson (2019).…”

Section: Discussioncontrasting

confidence: 56%

“…Studies applying feature selection methods cannot necessarily decrease the uncertainty regarding the choice of features. Malik et al (2023) found the set of chosen features depends strongly on the feature selection method. They also found that removing some features deemed by feature selection methods as less important is not necessarily improving but rather decreasing training accuracy.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

Edrich,

Yildiz,

Roscher

et al. 2023

Preprint

View full text Add to dashboard Cite

Machine learning (ML) has become increasingly popular in recent years for susceptibility and hazard mapping tasks. Necessary steps of ML-based workflows for various purposes in the context of geohazards are implemented repeatedly for new studies. To facilitate an effective, modular and extensible application of ML for susceptibility and hazard mapping, a generic framework based on a Random Forest (RF) classifier is presented herein. It supports especially data pre-processing and RF input dataset generation, as well as the susceptibility or hazard map generation adapted to an individual research question. The FAIR data principles were adopted as a guideline during the design and implementation of the framework. In addition to the reproducibility of the resulting binary maps, the implementation enables extensibility and scalability. The framework is applied to the test case of rainfall triggered shallow landslide susceptibility in canton Grisons, Switzerland. Computational experiments investigating the influence of the composition of the training dataset on the mapping result are conducted. Thereby the combination of features, sampling strategies for absence (non-landslide) locations, and transferability from local to regional scale were of interest. The results emphasise the importance of considering the composition of training data for the quality of the model and as a result also the reliability of the generated map.

show abstract

Section: Discussionsupporting

confidence: 87%

Section: Discussioncontrasting

confidence: 56%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

Edrich,

Yildiz,

Roscher

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…This is supported by the őndings in Liu et al (2021a) that feature selection is problem speciőc also with regard to the applied ML algorithm. While Nirbhav et al (2023) describe that feature selection methods might decrease the accuracy of the resulting model, Kuhn and Johnson (2019) showed for the RF a decrease in accuracy for only a large amount of added irrelevant features. The number for which this decrease was observed exceeds by far the number of features typically included for ML-based landslide susceptibility and hazard mapping.…”

Section: Discussionmentioning

confidence: 99%

“…Studies investigating feature selection methods claim that the quality of mapping results can be increased by dropping irrelevant features (e.g. Pham et al 2021;Nirbhav et al 2023). Nirbhav et al (2023) found that the set of chosen features depends strongly on the feature selection method.…”

Section: Discussionmentioning

confidence: 99%

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

Edrich,

Yildiz,

Roscher

et al. 2024

Preprint

View full text Add to dashboard Cite

Machine learning has grown in popularity in the past few years for susceptibility and hazard mapping tasks. Necessary steps for the generation of a susceptibility or hazard map are repeatedly implemented in new studies. We present a Random Forest classifier-based landslide susceptibility and hazard mapping framework to facilitate future mapping studies using machine learning. The framework, as a piece of software, follows the FAIR paradigm, and hence is set up as a transparent, reproducible and modularly extensible workflow. It contains pre-implemented steps from conceptualisation to map generation, such as the generation of input datasets. The framework can be applied to different areas of interest using different environmental features and is also flexible in terms of the desired scale and resolution of the final map. To demonstrate the functionality and validity of the framework, and to explore the challenges and limitations of Random Forest-based susceptibility and hazard mapping, we apply the framework to a test case. This test case conveys the influence of the training dataset on the generated susceptibility maps in terms of feature combination, influence of non-landslide instances and representativeness of the training data with respect to the area of interest. A comparison of the test case results with the literature shows that the framework works reliably. Furthermore, the results obtained in this study complement the findings of previous studies that demonstrate the sensitivity of the training process to the training data, particularly in terms of its representativeness.

show abstract

Towards Artificially Intelligent Landslide Susceptibility Mapping: A Critical Review and Open Questions

Teke,

Kavzoglu

2024

Advances in Natural and Technological Hazards Research

View full text Add to dashboard Cite

Landslide Susceptibility Prediction based on Decision Tree and Feature Selection Methods

Cited by 11 publications

References 66 publications

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

Towards Artificially Intelligent Landslide Susceptibility Mapping: A Critical Review and Open Questions

Contact Info

Product

Resources

About