Top-k Probability Of Frequent Co-location Pattern Algorithm Optimization And Database Implementations

Spatial co-location pattern represents a set of spatial features whose instances appear frequently in a spatial neighborhood. Spatial co-location pattern mining is an important filed, there are a lot of algorithms about certain dataset, but the task about spatial co-location pattern mining in uncertain dataset is in its infancy. Uncertain dataset exists everywhere in the real-life scenarios and the co-location mining algorithms in certain dataset cannot be applied to uncertain data, so spatial co-location pattern mining in uncertain dataset is very significant. This paper studies the top-k probabilistically prevalent co-location mining in uncertain dataset and the implement in database, as follows:First, we summarize the research of the spatial co-location pattern mining introduce the knowledge about uncertain dataset and the related definitions about spatial co-location pattern mining.Second, introduce the possible worlds, top-k probabilistically prevalent co-location and matrix method. By the analysis the characteristics of the matrix, we promote a method to get the participation based on index, improve the structure of the matrix, we propose two branches to reach one judgment-bulk pruning effect of possible worlds. At the same time, a grid method to get the table-instance.Third, we design a database application system about top-k probabilistically prevalent co-location mining, introduce the design of the table and stored procedure in detail.Fourth, we take experiments on synthetic and real dataset to verify the efficiency of the optimal algorithm and analyze the impact of various parameters and dataset distribution on the performance of the algorithm, the advantages and disadvantages of the database system.Finally, the summary of this paper and future work were presented.