Research of mining partial periodic co-occurrence patterns

With the development of modern information and the increase of space-time sensor and resolution, the volume of temporal and spatial data continues to be a significant grow. Therefore, mining valuable information from spatial and temporal data sets becomes more and more meaningful. The mining of periodic spatial and temporal co-occurrence pattern has been the central issue of recent research. The traditional algorithms for mining spatial temporal co-occurrence patterns are very time-consuming and having some redundant computation. On the other hand, these algorithms are based on threshold method. As we all know, the selection of threshold is suffering and lack of scientific basis. Therefore, T-PPCOP miner is proposed, which integrated TOP-K% method into the above algorithms to replace the confidence threshold method. Experimental results by real data sets show that the proposed T-PPCOP miner is feasible, and can effectively dig up partially periodic spatial temporal co-occurrence pattern (PPCOP) from spatial temporal data sets.