首先发展了基于事件影响域的时空事务表构建策略, 提出了基于事件影响域的时空关联规则挖掘方法, 给出了相应的挖掘算法(简称ECSTAR算法)。通过一个实际算例验证了所提方法的可行性和有效性。

Spatio-temporal association rules mining is a key technology and a hot issue in the field of spatio-temporal data mining. The classical Apriori algorithm is usually utilized to detect the spatio-temporal association rules from the spatio-temporal transaction table, which is derived from the original spatio-temporal data. In most existing approaches to generate the spatio-temporal transaction table, many defects, such as data redundancy, further affects the efficiency of spatio-temporal association rules mining. This paper proposes an events-coverage based spatio-temporal association rules mining (ECSTAR for short) to overcome these limitations. ECSTAR employs the event’s coverage to divide the researching spatio-temporal domain into some cells to generate a spatio-temporal transaction. Among each cell, spatio-temporal relationship predications are utilized to present the spatio-temporal relationship between the events and spatio-temporal objects. Thus, the spatio-temporal transaction table is built and spatio-temporal association rules are mined by the Apriori algorithm. Moreover, many concepts about ECSTAR are expounded and its algorithm is narrated in detail. Finally, a practical experiment demonstrates the feasibility and validity of the ECSTAR.