TY - JOUR
T1 - Bridge Deterioration Knowledge Ontology for Supporting Bridge Document Analytics
AU - Liu, Kaijian
AU - El-Gohary, Nora
N1 - Publisher Copyright:
© 2022 American Society of Civil Engineers.
PY - 2022/6/1
Y1 - 2022/6/1
N2 - Bridge owners possess important data sources, such as bridge construction records and inspection and maintenance reports, which hold great promise for improving understanding of bridge deterioration and informing maintenance decision making. However, the valuable data buried in these documents are not being fully exploited due to their unstructured nature. Domain-specific semantics are needed to facilitate analysis of the data based on content and domain-specific meaning and to bridge terminology gaps across different sources. To address this need, this paper proposes a bridge ontology (BridgeOnto) that captures deterioration knowledge and semantics related to bridge elements, deficiencies, deficiency causes, and maintenance actions. The proposed ontology was evaluated through answering competency questions, automated consistency and redundancy checking, expert interviews, and application-oriented validation. The BridgeOnto was implemented in supporting automated extraction of information describing bridge conditions and maintenance actions from bridge inspection reports. The experimental results show that the ontology was able to improve information extraction precision, recall, and F-1 measure by 11.7%, 12.4%, and 12.0%, on average. This research contributes to the body of knowledge by offering an ontology that can capture the key deterioration knowledge areas with sufficient coverage and in-depth classification for achieving adequate support for bridge document analytics. By allowing better access to and use of crucial textual data, this research has the potential to improve the understanding of bridge deterioration and enhance maintenance decisions.
AB - Bridge owners possess important data sources, such as bridge construction records and inspection and maintenance reports, which hold great promise for improving understanding of bridge deterioration and informing maintenance decision making. However, the valuable data buried in these documents are not being fully exploited due to their unstructured nature. Domain-specific semantics are needed to facilitate analysis of the data based on content and domain-specific meaning and to bridge terminology gaps across different sources. To address this need, this paper proposes a bridge ontology (BridgeOnto) that captures deterioration knowledge and semantics related to bridge elements, deficiencies, deficiency causes, and maintenance actions. The proposed ontology was evaluated through answering competency questions, automated consistency and redundancy checking, expert interviews, and application-oriented validation. The BridgeOnto was implemented in supporting automated extraction of information describing bridge conditions and maintenance actions from bridge inspection reports. The experimental results show that the ontology was able to improve information extraction precision, recall, and F-1 measure by 11.7%, 12.4%, and 12.0%, on average. This research contributes to the body of knowledge by offering an ontology that can capture the key deterioration knowledge areas with sufficient coverage and in-depth classification for achieving adequate support for bridge document analytics. By allowing better access to and use of crucial textual data, this research has the potential to improve the understanding of bridge deterioration and enhance maintenance decisions.
KW - Bridge data analytics
KW - Bridge deterioration
KW - Information extraction
KW - Ontology
KW - Semantic modeling
UR - http://www.scopus.com/inward/record.url?scp=85127913854&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85127913854&partnerID=8YFLogxK
U2 - 10.1061/(ASCE)CO.1943-7862.0002210
DO - 10.1061/(ASCE)CO.1943-7862.0002210
M3 - Article
AN - SCOPUS:85127913854
SN - 0733-9364
VL - 148
JO - Journal of Construction Engineering and Management
JF - Journal of Construction Engineering and Management
IS - 6
M1 - 04022030
ER -