성과구성 |
ABSTRACT
This paper aims to describe an approach to word embedding-based semantic analysis of building rule checking-related sentences for BIM-enabled automated rule checking system. Automated rule checking has been developed in various ways and enhanced the efficiency of building design review process. Converting human-readable building regulations to computerreadable format is, however, still time-consuming and error-prone due to the nature of human languages. The computer has to understand semantic meaning of regulatory sentences, in order to intelligently automate rule interpretation process. In this regards, this paper proposes a word embedding-based semantic analysis process of regulatory sentences and its utilization for composing rule set. The proposed semantic analysis process is composed of following steps: 1) preprocessing of building code sentences, 2) learning semantics of words and sentences with word embedding technique, 3) utilization for composing rule set. By utilizing word embedding technique, computer can learn semantic meaning of natural language and represent it in numerical values. In this paper, 3504 building code sentences are represented with numerical values, and those are used for classifying the topic of sentences. The results of learning are utilized for semantic searching system which help users to find out related regulations and compose rule sets.
Key Words: Automated rule checking, Building information modeling (BIM), Natural language processing, Semantic analysis, Word embedding |