Co - The Location With Time Constraint Pattern Mining

Along with the extensive application of spatial location, spatial data mining becomes one of the most promising research directions in the data mining, and co-location pattern mining is to find a collection of spatial objects of a set of frequently occurring, is the most important the research content in spatial data mining. Co-location pattern mining has been a considerable number of research results, however, these results which are based on the idea of equally instance does not adapt to all possible situations, the current study did not take into account an important factor of the time. In this paper, about the co-location model with time constraint, we advance the relevant concepts and algorithms.First, this paper introduces the basic concepts of spatial data mining and related definitions, characteristic, classification and research.Secondly, we introduce the co-location pattern mining, related concepts, and method of co-location pattern mining with rare characteristics, and describe the current status quo. Then this paper introduces the three algorithms:Join-Based algorithm which is used to co-location mining in certain data; Ujoin-based algorithm in uncertain data, and the top-k closed co-location pattern algorithm.Third, for the co-location model with time constraint, this article gives the relevant definitions and theorems, and proofs the theorem, then two basic algorithms are given: TCjoin-based algorithm, co-location pattern mining algorithm, top-k co-location pattern mining with time constraint algorithm. The former is suitable for spatial data mining background of personnel, and the latter to adapt to the personnel with no knowledge, do not give the minimum participation threshold. For these two algorithms, the paper gives optimization algorithms, in order to improve the time efficiency of the algorithm.Fourth, we make a lot of experiments through the simulation data, and analysis the impact of various parameters of the algorithm in order to arrive at the impact of various parameters on the algorithm, then compares optimization algorithms and the basic algorithm, It is proved that the optimization strategy is correct and effective. Finally, the conclusion makes a brief review of the contents of this article, and points out that the current co-location pattern mining research work is not completed and defects, and future research are presented.