| 张润梅,杨超,尹蕾,张媛.徽派建筑知识图谱的半自动化构建[J].安徽建筑大学学报,2021,29(): |
| 徽派建筑知识图谱的半自动化构建 |
| Semi-automatic Construction of Huizhou Architectural Knowledge Graph |
| 投稿时间：2020-12-21 修订日期：2021-04-08 |
| DOI： |
| 中文关键词: 徽派建筑 知识图谱 图数据库 BiLSTM-CRF模型 Neo4j |
| 英文关键词: huizhou architecture knowledge graph graph database BiLSTM-CRF model Neo4j. |
| 基金项目:安徽省自然科学基金(2008085MF218)；安徽省高校学科拔尖人才学术资助项目(gxbjZD26)；安徽省高校省级人文社会科学研究项目(SK2018A0578)；安徽省自然科学基金青年项目(1508085QF137)；校级科研项目(JZ192066)；校引进人才及博士启动基金项目(2019QDZ38)；安徽省教育厅自然科学项目（KJ2019JD12）。 |
|
| 摘要点击次数: 1843 |
| 全文下载次数: 569 |
| 中文摘要: |
| 摘要：针对徽派建筑领域知识智能化推荐和搜索系统需要构建专业知识库的问题，提出了一种BiLSTM-CRF模型与徽派建筑词典相结合的命名实体识别方法，利用先验知识的辅助作用，提升了BiLSTM-CRF模型的实体识别效果，完成了对建筑实体进行的识别抽取。利用Neo4j图数据库支持的属性图模型对知识进行表示。最后使用Neo4j图数据库存储知识，并在此基础上对构建的徽派建筑知识图谱进行了可视化展示。研究结果表明，此方法能够有效地构建徽派建筑领域知识图谱，为今后徽派建筑知识在智能化推荐和搜索系统的研究奠定基础。 |
| 英文摘要: |
| Abstract：Aiming at the problem that the intelligent recommendation and search system of Huizhou architectural field knowledge needs to build a professional knowledge base, this paper proposes a named entity identification method combining BiLSTM-CRF model with Huizhou architectural dictionary, using the auxiliary function of prior knowledge, the entity recognition effect of BiLSTM-CRF model is improved, and the recognition and extraction of architectural entities are completed. The knowledge is represented using the property graph model supported by the Neo4j graph database. After that, the Neo4j graph database is used to store knowledge and the Huizhou architectural knowledge graph is visualized on this basis. The research results show that this method can effectively construct a knowledge map in the field of Hui-style architecture, which lays a foundation for the future research of Hui-style architecture knowledge in intelligent recommendation and search systems. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
