Perdition of gully erosion susceptibility mapping using novel ensemble machine learning algorithms. Issue 1 (1st January 2021)

• Record Type: Journal Article
• Authors: Arabameri, Alireza; Chandra Pal, Subodh; Costache, Romulus; Saha, Asish; Rezaie, Fatemeh; Seyed Danesh, Amir; Pradhan, Biswajeet; Lee, Saro; Hoang, Nhat-Duc
• Is Part Of: Geomatics, natural hazards & risk. Volume 12:Issue 1(2021)
• Page Start: 469
• Page End: 498
• Publication Date: 2021-01-01
• Subjects: Gully erosion -- spatial modelling -- ensemble learning -- extreme gradient boosting machine -- GIS -- genetic algorithm; Geomatics -- Periodicals; Geomatics; Periodicals; 526.905
• DOI: 10.1080/19475705.2021.1880977
• Languages: English
• ISSNs: 1947-5705

Abstract

Abstract: Spatial modelling of gully erosion at regional level is very relevant for local authorities to establish successful counter-measures and to change land-use planning. This work is exploring and researching the potential of a genetic algorithm-extreme gradient boosting (GE-XGBoost) hybrid computer education solution for spatial mapping of the susceptibility of gully erosion. The new machine learning approach is to combine the extreme gradient boosting machine (XGBoost) and the genetic algorithm (GA). The GA metaheuristic is being used to improve the efficiency of the XGBoost classification approach. A GIS database has been developed that contains recorded instances of gully erosion incidents and 18 conditioning variables. These parameters are used as predictive variables used to assess the condition of non-erosion or erosion in a given region within the Kohpayeh-Sagzi River Watershed research area in Iran. Exploratory results indicate that the proposed GE-XGBoost model is superior to the other benchmark solution with the desired predictive precision (89.56%). Therefore, the newly built model may be a promising method for large-scale mapping of gully erosion susceptibility.