Spatial co-location pattern ordering

Abstract

Mining spatial co-location pattern is one of the most important researches in the field of spatial data mining. In the past researches, many spatial co-location pattern mining algorithms and the expansions about these algorithms have been proposed. However, some of these methods often produce a large number of patterns which are difficult to use. If we want to use the subset of the prevalent co-location pattern set to summarize the whole set and as the increase of the number of patterns in subset, the patterns in subset always are the best summary for the original prevalent set. This is a NP-hard problem. In this paper, we consider the problem of ordering a prevalent co-location pattern set so that each prefix of the ordering gives as good a summary of the set as possible. And according to the features of spatial data, we define an estimation of participation index function and a prevalent co-location pattern loss function to formulate this problem and design a greedy algorithm which gives an approximation quality.