Analysis on association rules of multi-field information of Baishuihe landslide based on the data mining

Abstract

In order to explore the association criteria of landslide multi-field monitoring data, we have adopted the two-step clustering method and Apriori algorithm, which belong to the classical data mining method, and we have also proposed the process of landslide monitoring data mining. Based on the Baishuihe landslide in the Three Gorges Reservoir Area, we analyzed the monitoring data of ZG93 from June 2003 to June 2016. The main inducing factors of the landslide displacement were selected, and the two-step clustering method was used to pre-cluster and cluster the different influence factors. We used Apriori algorithm to deal with the classified variables to generate frequent item sets that satisfy the minimum support degree. The association rules between the precipitating factors and the landslide deformation are established under the multi-field coupling mode of Baishuihe landslide. The results show that the correlation criterion is of great significance to the deformation analysis of landslide hazards and the data mining technology can be applied to the displacement prediction of geological hazards in the Three Gorges Reservoir Area.

Keywords

• reservoir landslide
• data mining
• two-step clustering method
• apriori algorithm
• association rules