Based On Pattern Fusion Mining Frequent Co-location Long Mode

Along with the development of geographic information system technology is increasingly perfect, and spatial data mining has become a hot research direction in data mining.one of the key is found co-location patterns in the space. A prevalent spatial co-location pattern is referred to as the frequent adjacency of spatial objects’ instance distribution within this pattern. In the past, lots of algorithms for mining co-location patterns have been proposed. However, the existing mining algorithms are of low efficiency in the very large spatial data sets, especially for mining spatial long co-location pattern. Aiming at the solution of massive data and long pattern mining problem, a pattern fusion based method for mining spatial long co-location pattern has been proposed in this paper.Firstly, introduced long mode defined in the traditional database and the long pattern mining research on status and significance.Secondly, it introduces the basic concepts of traditional co-location model and some classic mining algorithms. For example, join-based algorithm based on fully connected, some partial-join algorithm is based on the connection and so on.Thirdly, related concepts such as the co-location core pattern and co-location pattern distance are defined. The long co-location pattern mining algorithms based on the pattern fusion are presented based on this method.Fourthly, according to a long pattern mining algorithm, we proposed a co-location mining algorithms to the great mining long patterns further.Fifthly, a large number of experiments are made to prove the accuracy and efficiency of the proposed algorithms.Finally, this paper summarizes the work, study co-location space and the prospects for long patterns were discussed.