Mining frequent items in spatio-temporal databases

It is important to retrieve aggregate information in spatio-temporal applications. Recently, some applications, such as decision support systems, also require to mine frequent items based on a dataset within a query region during a query interval. Because of unbounded space requirement and slow response time, executing query based on operational databases becomes inapplicable. In this paper, we define the problem formally and give out a novel solution to overcome the above two disadvantages. Recently, some algorithms are proposed to mine frequent items from a summarization(sketch) of a mass dataset. In our solution, one of latest sketches is integrated with a spatio-temporal index to provide good performance.