Intelligent prediction and analysis of influencing factors of Quaternary accumulation layer thickness in landslide-prone areas: A case study in the Tiefeng area of Wanzhou District, Chongqing City

Accumulation layer thickness is the basic data of regional engineering geological investigation, which is of great significance to identifying landslides and plays an important role in landslide stability evaluation and risk assessment. The traditional interpolation method does not take into account external factors, and it is difficult to meet the accuracy requirements. This study takes the Tiefeng Township of Wanzhou District in the Three Gorges Reservoir as the research object, summarizing the formation model of the accumulation layer and regional accumulation layer thickness control points are added accordingly. The Apriori algorithm is used to mine the association rules between the influence factors and the thickness distribution of accumulation layers. Three machine learning methods are used to build the model by using the known sample points, which are applied to the entire area to obtain the thickness distribution map of the accumulation layer, then compare the results. Results show that 6 of the 13 selected influence factors have a strong correlation with the accumulation layer thickness, and slope and topographic undulation are the main factors that cause the difference in the spatial distribution of accumulation layer thickness. Among the three machine learning methods, the prediction result of the GWO-SVM model is the most consistent with the actual situation. The results reveal the formation mechanism of regional quaternary accumulation layer and lay a foundation for machine learning in the field of accumulation layer thickness prediction.