空间约束地址模型及推理匹配方法研究

Research on Spatial Constrained Address Model and Reasoning Matching Method

复杂多变的中文地址结构使得在进行地址匹配时存在一定的局限性,影响地址匹配效率。针对该问题,研究了适合空间推理的空间约束地址模型。在分析中文地址的结构特点以及地址描述中存在的问题的基础上,以地址要素间的空间关系为切入点,构建该模型。进行推理匹配时,首先借助空间约束关系进行推理识别;然后,采用深度优先、集合查找的策略,同时正向推理所在城市作为集合查找左边界;最后,逆向推理最小目标地址要素直到获取最优目标地址。以中文地址作为实验数据进行推理匹配的可行性验证。试验结果表明,顾及空间关系的推理匹配方法可以有效解决地址描述的多样性和空间约束不一致性等问题,极大地提高了检索效率和匹配成功率。

The complex and variable Chinese address structure makes certain limitations in address matching and affects address matching efficiency.In view of the above problems,this paper studies the spatial constrained address model suitable for spatial reasoning.Based on the analysis of the structural characteristics of the Chinese address and the problems in the address description,the model is constructed by taking the spatial relationship between the address elements as the entry point.When reasoning matching,the spatial constraint is used for reasoning recognition,first,the depth first search and set strategy is found by the left boundary.Finally reverse reasoning minimum target address elements until obtaining the optimal target address.This paper takes a random sample of 3500 Chinese address as the experimental data reasoning,to verify the feasibility.The test results show that,considering the spatial relation reasoning matching method can effectively solve the diversity and space constraints address description inconsistency problem,greatly improves the efficiency of retrieval and matching success rate.