대한건축학회논문집 (Journal of the Architectural Institute of Korea)

  • 제40권9호
  • /
  • Pages.247-254
  • /
  • 2024
  • /
  • 2733-6239(pISSN)
  • /
  • 2733-6247(eISSN)
대한건축학회 (Architectural Institute of Korea)
권호 표지
DOI QR코드

DOI QR Code

AutoML 기반의 이진, 다중 분류 모델 구축을 통한 건설 사고 발생 및 유형 예측

Predicting Construction Safety Accidents and Types Using AutoML-based Binary and Multi-class Classification Models

  • 전정호 (부산대학교 건축공학과 ) ;
  • 최수연 (부산대학교 건축공학과) ;
  • 윤성배 (부산대학교 건축공학과 ) ;
  • 용선진 (부산대학교 건축공학과) ;
  • 허영기 (부산대학교 건축공학과 )
  • Jeon, JungHo (Dept. of Architectural Engineering, Pusan National University) ;
  • Choi, SuYeon (Dept. of Architectural Engineering, Pusan National University) ;
  • Yun, SeongBae (Dept. of Architectural Engineering, Pusan National University) ;
  • Yong, SunJin (Dept. of Architectural Engineering, Pusan National University) ;
  • Huh, YoungKi (Dept. of Architectural Engineering, Pusan National University)
  • 투고 : 2024.07.22
  • 심사 : 2024.08.22
  • 발행 : 2024.09.30
⟨ 이전 논문 다음 논문 ⟩

초록

In 2022, the construction industry accounted for nearly half of all fatal accidents across sectors in South Korea. This study aims to develop binary like fatality and injury and multi-class such as fall and struck-by classification models using AutoML to predict the occurrence and types of construction accidents, based on data from the Construction Safety Integrated Management System (CSI) database. The dataset, consisting of 235,665 accident cases from January 2019 to February 2024, includes 54 types of information, with 18 influential accident factors identified. Preprocessed data were trained and tested using AutoML to determine optimal algorithms and influencing factors. Accuracy, precision, recall, and F1 score metrics were used for validation. The binary classification model for predicting fatalities and injuries, developed using the Extra Trees (ET) algorithm, achieved the highest accuracy of 95.9% and an F1 score of 0.2771. For predicting accident types, the multi-class classification model using the LightGBM (LGBM) algorithm recorded the highest accuracy of 57.4% and an F1 score of 0.5503. Feature importance analysis revealed that the accident object was the most critical factor in both models. This research is expected to enhance safety management performance by efficiently identifying the likelihood and types of construction accidents.

키워드

과제정보

이 과제는 부산대학교 기본연구지원사업(2년)에 의하여 연구되었음.

참고문헌

  1. Cho, M. G., Lee, D. G., Park, J. Y., & Park, S. H. (2021). Development of Machine Learning-based Construction Accident Prediction Model Using Structured and Unstructured Data of Construction Sites, Journal of Civil and Environmental Engineering Research, 42(1), 127-134.
  2. Cho, Y. R., Kim. Y. C., & Shin, Y. S. (2017). Prediction Model of Construction Safety Accidents using Decision Tree Technique, Journal of the Korea Institute of Building Construction, 17(3), 29.
  3. Choi, J. K. (2019). A Prediction Model for Fatal Accidents among Construction Workers using Machine Learning, MS thesis, Sungkyunkwan Univ., Korea.
  4. Choi, S. J., Kim, J. H., & Jung, K. H. (2021). Development of Prediction Models for Fatal Accidents using Proactive Information in Construction Sites, Journal of the Korean of Safety, 36(3), 31.
  5. Hwang, S. H., Kang, S. W., & Shin, Y. S. (2022). Study on the Risk Assessment of Fall Accidents at Construction Sites Through The Analysis of Accident Cases, Journal of the Korea Academia-Industrial cooperation Society, 23(12), 41-49.
  6. Jeong, J. M., & Jeong, J. W. (2020). Development of Framework for Integrated Work-Risk Breakdown Structurebased on Fatal Incident Cases in Construction Industry, Korean Journal of Construction Engineering and Management, 21(3), 115, 11-19.
  7. Kerim K., Omer E., & Asli P. G. (2023). Prediction of construction accident outcomes based on an imbalanced dataset through integrated resampling techniques and machine learning methods, Engineering, Construction and Architectural Management, 30(9), 4486-4517.
  8. Kim, H. Y., Lee, J. S., & Jang, Y. Y. (2022). Analyzing Patterns of Multi-cause Accidents From KOSHA's Construction Injury Case Reports Utilizing Text Mining Methodology, Journal of the Architectural Institute of Korea, 38(4), 237-244
  9. Kim, T. H., Kim, K. N., & Lee, M. J. (2023). Development of Work Safety Management Platform based on Construction Accident Big Data Analysis, Journal of the Architectural Institute of Korea, 39(1), 331-336.
  10. Lee, K. Y. (2020). Association Rule Mining Approach to Extracting Relationships of Accident Factors in Construction Sites, MS thesis, Pukyong National Univ., Korea.
  11. Lee, S. K. (2018). A Study on the Trends of Construction Safety Accident in Unstructured Text Using Topic Modeling, Journal of the Korea Academia-Industrial cooperation Society, 19(10), 176-182.
  12. Park, H. P., & Han, J. K. (2019). Development of Risk Assesment Index for Construction Safety Using Statistical Data, Journal of the Korea Institute of Building Construction, 19(4), 361.