Accident Analysis for Construction Safety Using Latent Class Clustering and Artificial Neural Networks

Ayhan, Bilal Umut
Tokdemir, Onur Behzat
Despite many improvements in safety management, the construction industry still has the highest potential for occupational injuries including High Severe (HS) work events, which result in injuries or fatalities, and Low Severe (LS) work events, which cause near misses or nonserious injuries. The analysis of incidents is highly dependent on the quality of records. Problems in recording and the heterogeneity of incident data may create conflicts while analyzing the relationship between attributes. The objective of the study was to develop a novel model to predict the outcomes of construction incidents using Latent Class Clustering Analysis (LCCA) and Artificial Neural Networks (ANNs) and determine necessary preventative actions. ANN has been used for many years to investigate the nonlinear relation between attributes and generate a logic between them. Herein, ANN was used to perform severity analyses of incidents utilizing real data, which were collected from various construction sites anonymously. Many factors affect the performance of ANN, including the size of the input and the heterogeneity of data. LCCA was used to seek out better performance and accuracy in ANN applications by reducing the heterogeneity of the incidents. By applying LCCA, attributes that possess different probabilities were clustered together and put into the ANN model. Then, the study concluded by providing a necessary preventative measure according to the result of incidents forecasted in advance. The research has two significant contributions. First, the hybrid model revealed promising results as the performance of the ANN-based predictive model was enhanced by addressing the heterogeneity of data. Second, the study presented professionals with practical preventative actions to avoid construction incidents according to the results of prediction.


Construction labor productivity modeling with neural networks
Sönmez, Rifat (American Society of Civil Engineers (ASCE), 1998-11-01)
Construction labor productivity is affected by several factors. Modeling of construction labor productivity could be challenging when effects of multiple factors are considered simultaneously. In this paper a methodology based on the regression and neural network modeling techniques is presented for quantitative evaluation of the impact of multiple factors on productivity. The methodology is applied to develop productivity models for concrete pouring, formwork, and concrete finishing tasks, using data compi...
Integrated Approach to Overcome Shortcomings in Current Delay Analysis Practices
Birgönül, Mustafa Talat; Dikmen Toker, İrem (American Society of Civil Engineers (ASCE), 2015-04-01)
Many factors, such as unforeseeable events, managerial and financial problems of contractors, insufficient technical capacity of site teams of contractors or Consultants, and so on, may lead to delays in construction projects. Proper analysis of compensability and quantum of a delay event is of prime importance. Any delay analysis application can be considered as a result of the combination of contract documents, scheduler, record-keeping mechanism at the site, communication among project participants, dela...
Fuzzy Structural Equation Model to Assess Construction Site Safety Performance
Gunduz, Murat; Birgönül, Mustafa Talat; Ozdemir, Mustafa (American Society of Civil Engineers (ASCE), 2017-04-01)
The main goal of this study is to examine the relationships between determinants of safety performance of construction sites. To achieve this goal, 168 observable variables and 16 latent dimensions that affect safety performance of construction sites were collected through extensive literature reviews, expert opinions, and face-to-face interviews. A questionnaire form was developed and administered to construction professionals as a data-collection tool about the observable variables and latent dimensions. ...
Overlapping Lattice Modeling for concrete fracture simulations using sequentially linear analysis
Aydın, Beyazıt Bestami; Tuncay, Kağan; Binici, Barış (Wiley, 2018-04-01)
Modeling concrete fracture is important in order to uncover accurately the sources of distress which lead to the damage or failure of structures. Many different numerical approaches have been used in the past employing either a smeared or a discrete cracking approach. Those models have difficulty in capturing the local nature of cracking, as well as the direction of crack propagation. Lattice modeling and peridynamics (PD) are some of the more recent nonlocal fracture simulation tools which possess advantag...
An approach to investigate relationship between speed and safety on urban arterials
Ardiç Eminağa, Zerrin; Akyılmaz, M. Özdemir; Department of Civil Engineering (2008)
Traffic safety is an important problem in today’s world with increasing number of fatalities and injuries in traffic accidents. For the solution of this problem, determination of accident prone locations on a network and reasons behind is an essential step, which is studied to some extend via different traffic accident analyses in the literature. While major factors affecting accident risk, such as speed, congestion, infrastructural aspects are known, it is still very difficult to figure out the interaction...
Citation Formats
B. U. Ayhan and O. B. Tokdemir, “Accident Analysis for Construction Safety Using Latent Class Clustering and Artificial Neural Networks,” JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT, pp. 0–0, 2020, Accessed: 00, 2020. [Online]. Available: