A query expansion method for retrieving online BIM resources based on Industry Foundation Classes
A query expansion method for retrieving online BIM resources based on Industry Foundation Classes
复制标题
一种基于行业基础类的在线BIM资源检索扩展方法
DOI：
10.1016/j.autcon.2015.04.006
复制
发表时间：
2015-08-01
影响因子：
10.3
通讯作者：
Yong, Jun-Hai
中科院分区：
文献类型：
--
作者：
Gao, Ge;Liu, Yu-Shen;Yong, Jun-Hai
随着建筑信息建模（BIM）技术的迅速普及，建筑产品库等BIM资源在万维网上迅速增长。因此，这也增加了快速找到足够接近用户特定需求的有用BIM资源的难度。基于关键字的搜索方法由于其易用性而得到广泛应用，但由于特定BIM文档和查询中术语的语义模糊性，其搜索精度往往不令人满意。为了解决这个问题，我们开发了一个原型语义搜索引擎，名为BIMSeek，检索在线BIM资源。本文的主要工作包括以下两个部分。首先，基于行业基础类（IFC），这是一个主要的标准BIM，领域本体的编码BIM的特定知识到搜索引擎构建。使用本体，BIM文档中的术语可以消除歧义和索引。其次，结合本体和局部上下文分析技术，提出了一种查询自动扩展方法，以提高检索性能。实验结果表明，与传统的基于关键字的查询扩展方法和基于WordNet的查询扩展方法相比，该方法具有更好的性能。搜索引擎可在www.example.com上找到。(C)2015 Elsevier B.V.版权所有。
With the rapid popularity of Building Information Modeling (BIM) technology, BIM resources such as building product libraries are growing rapidly on the World Wide Web. As a result, this also increases the difficulty for quickly finding useful BIM resources that are sufficiently close to user's specific needs. Keyword-based search methods have been widely used due to their ease of use, but their search accuracy is often not satisfactory because of the semantic ambiguity of terminologies in BIM-specific documents and queries. To address this issue, we develop a prototype semantic search engine, named BIMSeek, for retrieving online BIM resources. The central work consists of two parts as follows. Firstly, based on Industry Foundation Classes (IFC) which is a major standard for BIM, a domain ontology is constructed for encoding BIM-specific knowledge into the search engine. Using the ontology, terminologies in BIM documents can be disambiguated and indexed. Secondly, by combining the ontology and local context analysis technique, an automatic query expansion method is presented for improving retrieval performance. Compared with traditional keyword-based methods and WordNet-based query expansion methods, the experimental results demonstrate that our method outperforms them. The search engine is available at http://cgcad.thss.tsinghua.edu.cn/liuyushen/ifcqe/. (C) 2015 Elsevier B.V. All rights reserved.