Research On Spatial Co-location Pattern Mining Method Based On Optimal Distance Attenuation Coefficient

Spatial co-location pattern mining can discover a set of spatial features that occur frequently side by side in a spatial system and is of great value in reflecting the spatial association rules among geographic features.To solve the problem that the existing spatial colocation pattern mining methods are less spatially interactive and the distance attenuation coefficient of spatial interaction intensity is usually taken as empirical value,this paper intends to improve the participation rate and frequency of spatial co-location pattern based on the spatial interaction intensity,introduce spatial interaction into spatial co-location pattern mining,and replace the empirical value by solving the optimal distance attenuation coefficient of spatial interaction intensity.In order to improve the accuracy and sensitivity of spatial co-location pattern mining,an improved spatial pattern mining method is constructed and applied in practice.The main research work in this paper includes:(1)A hesitant fuzzy set calculation framework for spatial interaction strength instance size is proposed to solve the calculation problem of spatial interaction strength instance size,and to provide a basis for transforming spatial interaction strength formula to solve its optimal distance decay coefficient.The scale of the representation instance is expressed comprehensively by the quantifiable multiple attribute values of the feature instance,and the attribute values are converted into the hesitant fuzzy elements by the way of the hesitant fuzzy set,and their hesitation fuzzy scores are calculated to quantify the scale of the instance.This computing framework can handle inconsistent number of instance attributes and provides a unified computing framework.(2)Particle swarm optimization(PSO)is introduced to solve the optimal distance decay coefficient of spatial interaction strength,in order to solve the problem of using the optimal value instead of the empirical value for the distance decay coefficient.Based on the form of distance attenuation coefficient in the formula of the Spatial Interactive gravity model,through formula transformation,the instance size obtained by the calculation framework of the hesitant fuzzy set is treated as the actual value,thus the solution of distance attenuation coefficient is converted to the optimization problem of the particle swarm algorithm,and the optimal distance attenuation coefficient is solved by using the instance size and distance.The calculation of the optimal distance decay factor provides the basis for constructing a more efficient spatial pattern mining method.(3)Based on the optimal distance decay coefficient of spatial interaction intensity,an improved spatial co-location pattern mining method(ds-method)is constructed.On the basis of obtaining the sample size of spatial interaction intensity and the optimal distance decay factor,an improved spatial co-location pattern mining method is constructed by defining the participation rate and frequency of spatial interaction intensity of the patterns.Compared with join-based,ns-method,nd-method and ds-method which takes empirical values,the ds-method in this paper has the highest accuracy and sensitivity.The best value of distance attenuation coefficient effectively improves the results of spatial pattern mining.(4)Introduce ds-method into the convenience analysis of living service facilities in Nanning concentration area.By combining the contribution value of the living service facilities to the convenience of residential areas with the results of ds-method spatial co-location pattern mining,the convenience rate,convenience index and distribution convenience results of the living service facilities in the study area are calculated and analyzed,which can provide reference for the construction and improvement of the convenient living circle in the study area.