Spatial Maximum Co-location Model Cloud Theory Of The Excavation

Mining spatial co-location pattern is currently facing many challenges. First, considerable uncertainty and fuzziness exist in the spatial data, such as there is large number of people in shopping malls, he is very tall. Second, it is difficult to accurately set the distance threshold of space adjacent relationship, and small changes in the distance threshold will have great influences on the number and the quality of frequent patterns. Third, the large amount of the spatial data in spatial database leads to the strikingly huge number of frequent patterns. At last, generating the table instance of co-location patterns takes most of the time in the traditional co-location pattern mining algorithms. To solve above problems, great efforts are made in this paper to study the approach of mining spatial maximal co-location patterns based on the cloud theory.First of all, we analyze the current situation of the research at home and abroad and challenge which spatial co-location pattern mining is facing.In addition, we design a more efficient method which is called as degree-based approach to solve the maximum clique problem, and make an empirical study to verify its completeness, correctness and efficiency on the synthetic data set and DIMACS benchmark figure.Moreover, through improving the degree-based approach to solve the maximal clique of spatial features, and introducing maximal frequent patterns instead of all the frequent patterns, we design the approach of mining spatial maximal co-location pattern, and verify the correctness and efficiency of this algorithm in the data set of three parallel rivers nature reserve in Yunnan province.Furthermore, based on the approach of mining spatial maximal co-location patterns, we solve the uncertainty and fuzziness existing in the spatial data by employing the cloud theory, and introduce the new concept of spatial neighborhood relation which reduces the effect of the distance threshold on the number and quality of frequent patterns. Besides, we design the approach of mining spatial maximal co-location patterns based on the cloud theory, and verify the advantage of the algorithm on the heavy metal pollution data set.The final part of this paper briefly summarizes this research and points out the extensible research direction in future.