Research On Spatial Co-location Pattern Mining Based On Hesitant Fuzzy Sets

A spatial co-location pattern refers to a set of spatial features whose the instances frequently appear together in a geographic space.As one of the important research directions in the spatial data mining,spatial co-location pattern mining aimed at finding the spatial features whose the instances are frequent co-locate in neighboring domain.The traditional spatial co-location pattern mining method is based on classical sets and the spatial position of its feature instances is determined.However,the spatial position is inevitably uncertain,so it is necessary to explore the mining method of spatial isomorphism on fuzzy sets.In order to solve the problems of the fuzzy uncertainty in the spatial positions of the characteristic instance and the spatial co-location pattern mining,in this paper,by combining the theory and method of hesitant fuzzy set to handle the properties of the data,and established the method of spatial co-location pattern mining based on hesitant fuzzy set,including the method of spatial co-location pattern mining based on the hesitant fuzzy participation degree in score and the method of spatial co-location pattern mining based on the hesitant fuzzy participation degree in position,its corresponding pruning strategy is discussed,and experimental analysis and verification.The research work of this paper mainly includes:(1)The basic theoretical research of the thesis.This paper analyzes the related theories and research status,including the analysis of the spatial features and spatial instances,spatial adjacent relationship,line instances and table instances,the participation rate and participation degree and frequent spatial co-location pattern;The calculation principle and the complete calculation process of the classical join-based algorithm for spatial co-location pattern mining are studied;Fuzzy set theory was studied and its related development theory,including the concept in fuzzy sets and the expression forms in fuzzy objects,including the concept in hesitant fuzzy sets and the expression forms in hesitant fuzzy objects,the point is to give the score function of the hesitating fuzzy set.(2)Proposing and establishing a spatial co-location pattern mining method based on the hesitant fuzzy participation degree in score.This paper combines the theory of hesitant fuzzy set with spatial co-location pattern mining,defines the spatial characteristics in hesitant fuzzy set and its examples,combines the calculation rules of hesitant fuzzy set and its score function,deals with the fuzziness of data by using the score function which consider decision risk and attitude of decision makers in hesitant fuzzy set,proposes and defines the hesitant fuzzy participation rate in score and the hesitant fuzzy participation degree in score,and then establishes the spatial co-location pattern mining method based on the hesitant fuzzy participation degree in score,and gives its corresponding implementation algorithm.Then the corresponding pruning strategy is given to improve the operational efficiency of the algorithm,and the correctness and feasibility of the algorithm are verified by experimental analysis and algorithm evaluation.(3)Proposing and establishing a spatial co-location pattern mining method based on the hesitant fuzzy participation degree in position.In this paper,the definition of the instance position in hesitant fuzzy set is given,the concept of the instance position in hesitant fuzzy set is distinguished in detail,the concept of the instance position in hesitant fuzzy set and its related expression are given.By scoring function in hesitant fuzzy sets,puts forward and defines space proximity in score,its relevant calculation rules are given,and then puts forward and defines the hesitant fuzzy participation rate in position and the hesitant fuzzy participation degree in position,establishes the spatial co-location pattern mining method based on the hesitant fuzzy participation degree in position and gives the corresponding algorithm and the corresponding pruning strategy.The effectiveness and feasibility of the proposed algorithm and strategy are verified by experimental analysis and algorithm evaluation.