Data mining of correlation between fire disturbance habitat factors and spider communities

Abstract

We would like to know the effects of fire disturbance habitat factors on the structure of spider communities in the Cangshan Mountain of Dali in Yunnan province. In this paper, the habitat data of burned sites at different restoration stages in Cangshan Mountain are taken as samples. Firstly, we constructed a habitat data model based on autoregressive integrated moving average model. And we predicted the change tendency of structure of spider communities by this model. Secondly, we used association rules, decision trees and variable importance analysis methods to investigate the correlation between habitat factors and structure of spider communities. The results showed that the vertical variation of the structure of spider communities in Cangshan Mountain was obviously. The restoration years changed the richness and population of spider species. Habitat factors affected the structure type, organization level and stability of spider communities.