“…The combination of SOM with different clustering methods can surpass single analysis methods and provide superior classification results, such as K-means clustering (Zhang et al, 2023b), hierarchical clustering (Kim et al, 2020) and fuzzy C-means clustering (Yao et al, 2022). Based on the classification results, conventional groundwater chemical characterization methods such as Piper diagrams, Gibbs diagrams, ion ratio analysis and mineral saturation indices were used to elucidate the hydrochemical characteristics and hydrochemical processes (Gao et al, 2022, Lu et al, 2023. The spatial distribution of water hydrochemical components based on inverse distance weight interpolation is essential to understand the distribution and evolution of groundwater contamination at spatial scales (Nsabimana and Li, 2023).…”