基于数据立方体的多最小支持度关联规则在犯罪分析中的应用

Application of Association Rules with Multiple Minimum Supports Based on the Data Cube in Crime Analysis

为了快速获取候选项集的支持度,避免频繁访问数据库而造成效率低下的问题,在MSapriori算法的基础上引入数据立方体,提出DC_MSapriori算法。该算法无需多次扫描事务数据库,减少了I/O操作,降低了搜索开销。实验基于福州市鼓楼区各大医院周边的案事件数据,快速挖掘出犯罪时空模式,验证了算法的有效性。

To achieve rapid acquisition of candidate set support degree and avoid low efficiency issue due to accessing frequently to the database, the DC_MSapriori algorithm based on introducing the data cube for the MSapriori algorithm is proposed in this paper. The corresponding experiments are implemented to quickly mine space-time crime patterns based on case data around hospitals in Gulou district of Fuzhou. In this way, the new algorithm characteristics on transaction database scanning, search cost and the I/O operations are revealed and the effective-ness of the algorithm is validated.