Abstract: 【Purpose/significance】In order to find more abundant and applicable counter-terrorism intelligence information,this paper
focuses on the mining method of multi-level association rule based on the single-level association rule mining.【Method/process】Ac?
cording to the data characteristics of counter-terrorism intelligence,a method of using uniform minimum support degree and multiple
single minimum support degree parameters is proposed to select frequent itemsets of multi-level terror-related features.In the mining
process,it should save some special redundant frequent itemsets,redundant multi-level association rules and uninteresting multi-level
association rules.【Result/conclusion】This method could find the correlation relationship of different concept hierarchy in the ter?
ror-related data.【Innovation/limitation】This method could make up the deficiency of the common single-level association rule min?
ing in the analysis of the terror-related data containing multi-layer attributes,and provide more scientific and abundant comprehen?
sive references for the counter-terrorism early warning and counter-terrorism decision-making.