Automated staff assignment for building maintenance using natural language processing

Yunjeong Mo, University of North Florida
Dong Zhao, Michigan State University
Jing Du, University of Florida
Matt Syal, Michigan State University
Azizan Aziz, Carnegie Mellon University
Heng Li, Hong Kong Polytechnic University

Abstract

Staff assignment is the decision-making to determine appropriate workforce with required skills to perform a specific task. Staff assignment is critical to success of construction projects, especially when responding to routine requests such as the change order and building service. However, the effectiveness is low due to manual processing by the management personnel. To improve the productivity of staff assignment, this paper creates a machine learning model that reads service request texts and automatically assigns workforce and priority through the technique of natural language processing (NLP). The dataset used for modeling in this study contains 82,106 building maintenance records for a 3-year period from over 60 buildings on a university campus. The results show a 77% accuracy for predicting workforce and an 88% accuracy for predicting priority, indicating a considerably high performance for multiclass and binary classifications. Different from existing studies, the NLP model highlights the value of stop-words and punctuation in learning service request texts. The NLP model presented in this study provides a solution for staff assignment and offers a piece of the puzzle to the information system automation in the construction industry. This study has an immediate implication for building maintenance; and, in the long term, contributes to human-building interactions in smart buildings by connecting human feedback to building control systems.

 

Accessibility Statement

This item was created or digitized before April 24, 2027, or is a reproduction of legacy material created before that date. It is preserved in its original, unmodified state specifically for research, reference, or historical recordkeeping. In accordance with the ADA Title II Final Rule, the Library provides accessible versions of archival materials by request. If you are experiencing difficulty accessing the information on the site due to a disability, please submit a request through the following form for assistance.