2019 CSCE Annual Conference - Laval (Greater Montreal)

2019 CSCE Annual Conference - Laval (Greater Montreal) Conference


Title
A Machine Learning-Based Approach for Building Code Requirement Hierarchy Extraction

Return to Session

Author(s)
Mr. Ruichuan Zhang, University of Illinois at Urbana-Champaign (Presenter)
Ms. Nora El-Gohary, University of Illinois at Urbana-Champaign
Abstract

To reduce the time, cost, and errors of compliance checking, various automated code compliance checking (ACC) methods have been developed and implemented. Although these methods have achieved different levels of automation, representativeness, and accuracy, most of them are unable to fully automatically convert complex building code requirements into computer processable forms. Such complex requirements usually have hierarchically complex clause and sentence structures (e.g., nested syntactic and semantic structures, conjunctive and alternative obligations, multiple exceptions, etc.). There is, thus, a need to decompose such complex requirements into hierarchies of much smaller, manageable requirement units that would be processable using most of the existing ACC methods. Rule-based methods have been used to deal with such complex requirements and have achieved high performance. However, they lack scalability, because the rules are developed manually and need to be updated and/or adapted when applied to a different type of building code. More research is, thus, needed to develop a scalable approach to automatically convert the complex requirements into hierarchies of requirement units to facilitate the succeeding steps of ACC such as deep information extraction, information transformation, data matching, and compliance reasoning. To address this need, this paper proposes a new, machine learning-based approach to automatically extract the aforementioned requirement hierarchies. The proposed approach consists of three primary tasks: (1) model the requirement hierarchies, requirement units, and relationships between the units; (2) use a machine learning-based approach to automatically segment the building code requirement sentences into requirement units; and (3) use natural language processing techniques to automatically detect the relationships between requirement units, both within a sentence and between sentences of multi-sentence requirements. The proposed approach was implemented and tested on a number of chapters from the 2009 International Building Code (IBC) and the Champaign 2015 IBC Amendments, and was evaluated using precision and recall. The preliminary results show that the proposed approach is able to extract the requirement hierarchies  from different types of complex requirements, with high precision and recall. The paper also discusses how the requirement hierarchies can be used in the succeeding steps of ACC such as deep information extraction.