项目名称: 基于进化算法的大规模本体匹配问题研究
项目编号: No.61503082
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 自动化技术、计算机技术
项目作者: 薛醒思
作者单位: 福建工程学院
项目金额: 20万元
中文摘要: 基于进化算法的本体匹配技术可以有效地解决小规模本体间的异质问题。然而,当本体中的实体规模非常大的时候(可能拥有几十万甚至上百万个实体),对于基于进化算法的本体匹配技术而言,如何减小进化算法的搜索空间、提高算法的搜索效率是有效解决大规模本体匹配问题的关键。本课题将在前期研究的基础上,深入分析大规模本体匹配过程的特点,重点研究面向本体匹配结果的大规模本体概念层的划分技术、基于进化算法的大规模本体的概念层和实例层的匹配技术。通过构建大规模本体匹配问题的优化模型、设计高语义识别度的相似度度量技术和综合应用元模型、并行策略和局部搜索策略的高效进化算法以获取高质量的大规模本体匹配结果,实现本体之间的交互与协作,促进本体工程、数据集成和智能语义检索等领域的发展。在相关方法和理论研究基础上,课题还将开发大规模本体匹配技术测试实验平台,并通过被广泛认可的测试数据验证大规模本体匹配技术的有效性。
中文关键词: 异构知识交互
英文摘要: Ontology matching technologies based on Evolutionary Algorithm (EA) are able to effectively solve the heterogeneous problem between small scale ontologies. However, when the scale of entities in the ontology becomes very large (probably tens of thousands of entities), for ontology matching approaches based on EA, it's critical to reduce the search space and improve the efficiency of the search. On the basis of the previous studies, this project will further analyze the characteristics of large scale ontology matching process, and mainly focus on the ontology alignment oriented ontology schema level partition technology, large scale ontology matching technologies based on EA on both schema level and instance level. Through the construction of optimization model of large scale ontology matching problem, design of similarity measures characterized by high semantic power and high efficient EA using meta-model, global parallel strategy and local search strategy to obtain high quality alignment, implement the cooperation between various ontologies, and finally improve the development of domain such as ontology engineering, data integration and intelligent semantic retrieval. Based on the related theories and methods, this project will also develop a large scale ontology matching technology test platform where the effectiveness of the large scale ontology matching technologies can be verified through widely accepted benchmark data.
英文关键词: Heterogeneous knowledge interaction