A Clustering Co - Location Pattern Mining Method Based On Mission

With the advent of the era of big data, mining useful information from massive data for decision support has been more widely used. Currently, mobile positioning technology has been widely used, mobile maps is increasingly becoming an indispensable tool for people to travel. Spatial data collection is getting easier. The information contained in the spatial data needs to be further tapped. As an important part of spatial data mining, spatial co-location pattern mining become more and more important. However, the current co-location space pattern mining is also facing many challenges. Mined too many prevalent co-location patterns, get a lot of co-location rules, it’s difficult for users to pick and choose the rules that are interested in is one of them. Closed prevalent co-location patterns are a representative of prevalent patterns and contain complete information about the prevalent co-location patterns. Therefore, mining closed prevalent co-location patterns can reduce the number of prevalent co-location patterns and co-location rules, let users find their Interest rules.This is a paper about closed co-location patterns mining. In this paper, I propose a new mining closed prevalent co-location patterns algorithm. The basic idea is to classify all the patterns according to their corresponding maximal instances clique set and participation. When mining closed prevalent co-location patterns, we count only the longest pattern of each class. It reduces the computation of the algorithm by reducing the number of co-location patterns that are generated. The basic steps of the algorithm are to find firstly maximal instances cliques that are used to mine closed event sets and calculate participations, then mine closed event sets and their lattice structure that are used to find out the longest pattern of each class and facilitate to judge whether the co-location patterns mined are closed later, and finally get the closed prevalent co-location patterns. In this paper, we prove the completeness of the algorithm, and analyze the efficiency of the algorithm by experiments. The extensive experiments demonstrate that this algorithm is an efficient algorithm when the minimum participation threshold is small or the parameter setting of minimum participation threshold is changed.