Regulatory information transformation ruleset expansion to support automated building code compliance checking

The traditional building code compliance checking process mainly relies on design reviewers to review documents or models manually. intensive manual effort needed makes this time-consuming, costly, and error-prone. Automated (ACC) could be a promising upgrade of the checking. With reduced workload reviewers, ACC is cheaper, faster, immune human errors. To support ACC, requirements need represented in computer-processable format enable automated reasoning, which will turn allow an assessment design's status with codes. A major limitation many existing systems/methods their limited range checkable requirements. address that, state art uses pattern matching-based rules transform into computable formats automatically, but ruleset was developed tested only few chapters An efficient expansion method increase its at low cost, bring systems closer full deployment. In paper, authors proposed new regulatory information transformation for expanding ruleset. This can expand without significant effort. takes iterative approach ensure generality validity quality results. expanded generating logic clauses from Chapter 5 International Building Code 2015. Compared baseline ruleset, increased predicate-level precision, recall, F1-score clause generation by 10.44%, 25.72%, 18.02%, 95.17%, 96.60%, 95.88%, respectively.