基于改进BP神经网络的基础地理信息与水文GIS信息语义映射研究

发布时间：2018-04-30 05:17

本文选题：基础地理信息本体 + 水文GIS　；参考：《东华理工大学》2017年硕士论文

【摘要】：随着我国科学新技术的兴起,我国城市形态也在不断的改革转型,数字城市向智慧城市成功升级。以数字城市为基础,依托“互联网+”平台,爆炸式的信息融入到我们的生活中。智慧城市将带来一场数据集成、共享、互操作的大变革,将各个行业的数据,包括行业的应用、行业的专题数据、行政审批数据,以及物联网的手机、视频、定位等等相关的数据往地理空间上面进行集成将是时下热点。数据呈现大数据的形态,面向越来越大的数据量,数据虽“大”但是并没有体现其应有的价值。不同的领域在各自信息化发展中都逐渐形成各自行业体系及相对独立的数据体系,对于同一内容,由于领域的不同导致数据概念有所差异。如何实现不同领域、不同系统间的数据互操作,以便实现数据的重用、语义映射是当前亟需解决的问题之一。在智慧城市的总体战略部署下,不同组织和结构下的信息共享与交流是智慧城市建设的数据基础,地理信息数据是用来实现地理智能服务的重要基础数据之一。随着本体语义和语义相似度的研究逐渐深入,给跨领域本体间语义映射问题的解决带来了新的思路。本文以基础地理信息与水文GIS本体的语义映射为研究对象,提出了一种面向地理实体的语义相似度计算与映射模型,利用改进的BP神经网络深度学习的方式来模拟人类主观评价概念相似度的思维过程,通过改进1)优化神经网络算法的全局搜索能力,力求寻找最优点,2)通过遗传操作中的选择算子、交叉算子、变异算子来优化神经网络算法每个神经元间的连接权值和阈值,3)解决不同本体非线性关系之间的映射问题。这些改进均进一步提升了传统语义相似度计算模式,并基于该模型实现对基础地理信息本体与水文GIS本体的跨领域的语义映射试验。实验结果显示,改进的算法精度优于传统BP算法,同时对改进模型进行评价,也证实该方式切实有效,可以辅助实现跨领域对象、实体的映射、融合、共享。
[Abstract]:With the rise of new science and technology in our country, the urban form of our country is constantly reforming and transforming, and the digital city has been upgraded successfully to the intelligent city. Based on the digital city, relying on the "Internet" platform, explosive information into our lives. Smart City will bring about a big change in data integration, sharing, and interoperability, bringing data from all sectors, including industry applications, industry thematic data, administrative approval data, mobile phones and video from the Internet of things. Positioning and other related data to the geospatial integration will be a hot spot. The data presents big data's form, facing the increasing amount of data, although the data is "large", it does not reflect its due value. Different fields have gradually formed their own industry systems and relatively independent data systems in the development of their respective information systems. For the same content, the different fields lead to different data concepts. How to realize data interoperability in different fields and systems in order to realize data reuse is one of the problems that need to be solved urgently. Under the overall strategic deployment of the intelligent city, the information sharing and communication under different organizations and structures is the data base of the construction of the intelligent city, and the geographic information data is one of the important basic data used to realize the geographic intelligent service. With the further research of ontology semantics and semantic similarity, a new way of thinking has been brought to the solution of semantic mapping problem between different domains. In this paper, the semantic mapping of basic geographic information and hydrological GIS ontology is taken as the research object, and a semantic similarity calculation and mapping model for geographic entities is proposed. Using the improved BP neural network depth learning method to simulate the human subjective evaluation concept similarity thinking process, by improving 1) optimize the global search ability of the neural network algorithm. We try to find the best point by optimizing the connection weight and threshold value of each neuron in the neural network algorithm by using the selection operator, crossover operator and mutation operator in the genetic operation) to solve the mapping problem between different ontology nonlinear relations. These improvements further enhance the traditional semantic similarity computing model and implement cross-domain semantic mapping experiments between the basic geographic information ontology and the hydrological GIS ontology based on the model. The experimental results show that the accuracy of the improved algorithm is better than that of the traditional BP algorithm. At the same time, the improved model is evaluated. It is also proved that this method is effective and can help to realize the mapping, fusion and sharing of cross-domain objects and entities.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P208

【参考文献】