Mining High-Utility Co-location Patterns From Spatial Datasets With Fuzzy Features

Co-location Pattern Mining(CPM)is based on large spatial datasets.The purpose of CPM is to search frequently associated objects in spatial neighborhoods.In the fields of biology,environmental science and epidemiology,it is very important to explore these models.However,CPM has great computational challenges,mainly due to the hidden association between schema instances in spatial data.In our daily life,vague objects can be seen everywhere,such as "young people" and"beautiful houses".It is precisely because there is no unique property of definite boundary between fuzzy objects and their relations that arouses the interest of a large number of scholars at home and abroad.In the field of spatial co-location pattern mining,previous studies have been limited to the fuzzy attributes and the threshold range of fuzziness,while ignoring the fuzziness features and the utility of such features themselves.This will result in a large number of low-frequency but valuable patterns not being discovered,while mining a large number of high-frequency patterns that users do not need.Therefore,for the first time,we introduce the concept of fuzzy features into spatial efficient co-location pattern mining to mine spatial efficient patterns satisfying the conditions of fuzzy features.In this way,on the one hand,it enlarges the scope of mining,and makes the information excavated more valuable;on the other hand,it greatly reduces unnecessary data and makes the efficiency of mining rapidly improve.In this paper,we apply the concept of fuzzy set theory to efficient co-location pattern mining,which can find efficient co-location patterns from fuzzy data sets.The main work is as follows:Firstly,the external utility and internal utility of fuzzy features are defined.Based on these two concepts,the concepts of fuzzy utility and fuzzy efficiency are further introduced.This paper designs the fuzzy efficiency threshold of the pattern to measure the spatial co-location pattern,and proposes an efficient spatial mining algorithm based on fuzzy features(FHUCB).Secondly,by constructing the attribute of weighted utility down-closing of star-shaped row instances,the fuzzy feature-based spatial efficient mining optimization algorithm based on co-location pattern(FHUCO)is constructed.Thirdly,the proposed algorithms are compared and analyzed from the aspects of mining result,mining efficiency,mining effect and so on the real data set and the synthetic data set.The experimental results show that the proposed method has good performance in terms of execution efficiency under various parameter settings.Finally,this paper also compares and analyzes the participation index and utility index between FHUCB algorithm and traditional co-location pattern mining algorithm,which shows the significance of this research.