See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/277011797 Pattern extraction for high-risk accidents in the construction industry: a data-mining approach Article in International Journal of Injury Control and Safety Promotion · September 2016 DOI: 10.1080/17457300.2015.1032979 CITATIONS 0 READS 31 4 authors, including: Some of the authors of this publication are also working on these related projects: Human Presence Detection using Innovative Sensing Approaches View project Mehran Amiri Amirkabir University of Technology 17 PUBLICATIONS 8 CITATIONS SEE PROFILE Elahe Soltanaghaei University of Virginia 5 PUBLICATIONS 1 CITATION SEE PROFILE All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately. Available from: Elahe Soltanaghaei Retrieved on: 03 October 2016
15
Embed
Pattern extraction for high-risk accidents in the construction industry: a data-mining ...es3ce/Elahe Soltan_files/papers... · 2016-10-03 · Pattern extraction for high-risk accidents
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
This article was downloaded by: [Abdollah ardeshir]On: 21 May 2015, At: 11:54Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK
Click for updates
International Journal of Injury Control and SafetyPromotionPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/nics20
Pattern extraction for high-risk accidents in theconstruction industry: a data-mining approachMehran Amiria, Abdollah Ardeshira, Mohammad Hossein, Fazel Zarandib & ElaheSoltanaghaeica Civil and Environmental Engineering Department, Amirkabir University of Technology,Tehran, Iranb Department of Industrial Engineering and Management Systems, Amirkabir University ofTechnology, Tehran, Iranc Computer Engineering Department, Sharif University of Technology, Tehran, IranPublished online: 21 May 2015.
To cite this article: Mehran Amiri, Abdollah Ardeshir, Mohammad Hossein, Fazel Zarandi & Elahe Soltanaghaei (2015): Patternextraction for high-risk accidents in the construction industry: a data-mining approach, International Journal of InjuryControl and Safety Promotion, DOI: 10.1080/17457300.2015.1032979
To link to this article: http://dx.doi.org/10.1080/17457300.2015.1032979
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
Pattern extraction for high-risk accidents in the construction industry: a data-mining approach
Mehran Amiria, Abdollah Ardeshira*, Mohammad Hossein Fazel Zarandib and Elahe Soltanaghaeic
aCivil and Environmental Engineering Department, Amirkabir University of Technology, Tehran, Iran; bDepartment of IndustrialEngineering and Management Systems, Amirkabir University of Technology, Tehran, Iran; cComputer Engineering Department,
Sharif University of Technology, Tehran, Iran
(Received 9 January 2014; accepted 10 December 2014)
Accidents involving falls and falling objects (group I) are highly frequent accidents in the construction industry. Whilebeing hit by a vehicle, electric shock, collapse in the excavation and fire or explosion accidents (group II) are much lessfrequent, they make up a considerable proportion of severe accidents. In this study, multiple-correspondence analysis,decision tree, ensembles of decision tree and association rules methods are employed to analyse a database of constructionaccidents throughout Iran between 2007 and 2011. The findings indicate that in group I, there is a significantcorrespondence among these variables: time of accident, place of accident, body part affected, final consequence ofaccident and lost workdays. Moreover, the frequency of accidents in the night shift is less than others, and the frequency ofinjury to the head, back, spine and limbs are more. In group II, the variables time of accident and body part affected aremostly related and the frequency of accidents among married and older workers is more than single and young workers.There was a higher frequency in the evening, night shifts and weekends. The results of this study are totally in line with theprevious research.
Keywords: pattern extraction; construction safety; high-risk accidents; multiple correspondence analysis; ensembles ofdecision tree; association rules
1. Introduction
Occupational accidents are the cause of more than
300,000 mortalities and 300 million injuries around the
world each year (International Labour Organization,
2013). This considerable number of cases has led to
severe human and financial impacts in different countries
(Warch, 2002). Previous studies have shown that workers
in various industries are vulnerable to occupational acci-
dents in different ways (Dudarev, Karnachev, & Odland,
2013). Construction is known as one of the most danger-
ous industries all over the world (Cheng, Leu, Lin, & Fan,
2010).
Occupational safety in the construction industry is
studied in different countries and regions around the
world (such as Halvani, Jafarinodoushan, Mirmoham-
On the other hand, the second group of accidents has
also been identified as serious and fatal accidents in
Table 5. Extracted association rules for circumstances and consequences of group I.
Association rule
ID Body part Head part Confidence (%) Lift
1 Lost workdays between 31 and 60 days Complete recovery 96 �2 Lost workdays between 31 and 60 days, and imprudence Complete recovery 96 �3 Age of worker between 20 and 35 Complete recovery 93 �4 Lost workdays between 61 and 120 days Disability between 10% and 33% 97 �5� Back or spine injury Age of worker between 20 and 35 64 1.23
6� Cranium and brain injury No lost workdays 57 1.14
7� Back or spine injury No lost workdays 44 1.12
�These rules are related to a part of accidents in group I which resulted in death or disability.
Table 6. Extracted association rules for circumstances and consequences of group II.
Association rule
ID Body part Head part Confidence (%)
1 Lost workdays between 1 and 30 days Complete recovery or disability between 10% and 33% 95
2 Lost workdays between 1 and 30 days, and age of workerbetween 20 and 35
Complete recovery or disability between 10% and 33% 99
3 Lost workdays between 61 and 120 days Complete recovery or disability between 10% and 33% 94
4 Lost workdays between 61 and 120 days, and imprudence Complete recovery or disability between 10% and 33% 96
5 Hand injury Complete recovery or disability between 10% and 33% 92
6 Limbs injury Complete recovery or disability between 10% and 33% 91
7 Accident occurrence in spring Complete recovery or disability between 10% and 33% 90
8 Age of worker between 20 and 35 days, and imprudence Complete recovery or disability between 10% and 33% 89
9 Imprudence Complete recovery or disability between 10% and 33% 89
10 Accident occurrence in summer Complete recovery or disability between 10% and 33% 88
11 Accident time between 7:01 and 10:00 am Complete recovery or disability between 10% and 33% 88
12 Accident time between 10:01 am and 12:30 pm Complete recovery or disability between 10% and 33% 90
13 Age of worker between 20 and 35 Complete recovery or disability between 10% and 33% 87
14 Accident occurrence in summer and imprudence Complete recovery or disability between 10% and 33% 91
15� Accident occurrence on Monday No lost workdays 83
16� During commuting to the workshop No lost workdays 82
17� Accident time between 10:01 am and 12:30 pm Inside the workshop 82
18� Accident occurrence in summer and inside the workshop No lost workdays 82
�These rules are related only to fatal accidents in group II.
International Journal of Injury Control and Safety Promotion 11
previous studies (Im et al., 2009; M€ungen & G€urcanli,2005; Scallan et al., 2004; Su�arez-Cebador, Rubio-
Romero, & L�opez-Arquillos, 2014). In this group, it was
observed that the frequency of accidents among married
workers is more than single ones. This finding is probably
due to the occurrence of these accidents among older
workers (which matches to the observations). Ling, Liu,
and Woo observed that the frequency of severe accidents
in the elderly is higher than in other age groups. They
associated the reason to repeating an activity in their work
and loss of consciousness in the elderly (Ling, Liu, &
Woo, 2009). According to the results, the frequency of
this group of accidents during lunch hours is higher than
other accidents. This result is in line with past research
(Su�arez-Cebador et al., 2014). It is also observed that the
frequency of occurrence of the second group of accidents
in the afternoon and especially night hours, and also on
weekends is much more than other accidents at work.
This may be due to executing earth-moving activity at
night and on weekends in Iran (to observe special urban
traffic provisions for soil moving machinery, etc.). More-
over, accidents that occurred outside the workshop or dur-
ing commuting to the workshop (which are mostly related
to being hit by vehicle accidents) are more frequent in this
group than others. Injuries to the head, face and neck in
this group are more frequent than other accidents that had
more severe (fatal and disabling) results. In this regard,
the ratio of accidents with no lost workdays (which are
probably related to instant death of a worker) and acci-
dents with more than 60 lost workdays (which are proba-
bly related to the disability of a worker) are also greater
than other accidents in the community studied.
Conducting the MCA technique, it was found that in
group II of accidents, there is a high correlation between
time of accident and body part injured. This finding is
implied in past research. For instance, Loudoun showed
that the time of accident has an impact on the severity
(Loudoun, 2010). On the other hand, a statistically signifi-
cant association between the injured body part and sever-
ity is also reported by Dumrak, Mostafa, Kamardeen, and
Rameezdeen (2013).
4.1. Limitations of the study
Archiving the attributes of occupational accident digitally
in the ISSO has just started about six years ago and is still
not in accordance with comprehensive classifications and
formats. In addition, the quality of gathering accident
information by work inspectors is not yet satisfactory.
Hence, some important variables such as worker occupa-
tion could not be considered in this study. Moreover,
although according to the Iranian law ISSO must be noti-
fied of all occupational accidents causing injury to insured
workers, it is possible that some cases remain unreported
or misreported. The ISSO does not archive the attributes
of near misses yet; therefore, this study is only based on
accidents happened. Despite these limitations, this study
was defined to be the first application of data-mining tech-
niques on the occupational accident data of Iran.
The results of this study confirm the results of previ-
ous studies as a whole; hence, it can be concluded that the
application of data-mining techniques has been success-
ful. In this regard, the capabilities of these techniques in
modelling large databases and detecting relationships
between variables were identified as their advantage.
Finally, the identified accident occurrence patterns can
assist policy-makers, managers and safety professionals in
the design and implementation of preventive measures
and strategies.
Investigating other serious types of accidents and also
analysing accidents considering weather conditions or
their geographical distribution in the country could be
considered as suitable subjects for future research.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Ale, B.J., Bellamy, L.J., Baksteen, H., Damen, M., Goossens, L.H., Hale, A.R., . . . Whiston, J.Y. (2008). Accidents in theconstruction industry in the Netherlands: An analysis ofaccident reports using storybuilder. Reliability Engineering& System Safety, 93(10), 1523�1533.
Amiri, M., Ardeshir, A., & Fazel Zarandi, M.H. (2014). Risk-based analysis of construction accidents in Iran during 2007-2011-meta analyze study. Iranian Journal of Public Health,43(4), 507�522.
Bevilacqua, M., Ciarapica, F.E., & Giacchetta, G. (2008). Indus-trial and occupational ergonomics in the petrochemical pro-cess industry: A regression trees approach. AccidentAnalysis & Prevention, 40(4), 1468�1479.
Brin, S., Motwani, R., & Silverstein, C. (1997). Beyond marketbaskets: Generalizing association rules to correlations. ACMSIGMOD Record, 26(2), 265�276.
Camino L�opez, M.A., Ritzel, D.O., Fontaneda, I., & Gonz�alezAlcantara, O.J. (2008). Construction industry accidents inSpain. Journal of Safety Research, 39(5), 497�507.
Chang, L.Y., & Wang, H.W. (2006). Analysis of traffic injuryseverity: An application of non-parametric classification treetechniques. Accident Analysis & Prevention, 38(5),1019�1027.
Cheng, C.W., Leu, S.S., Lin, C.C., & Fan, C. (2010). Character-istic analysis of occupational accidents at small constructionenterprises. Safety Science, 48(6), 698�707.
Cheng, C.W., Lin, C.C., & Leu, S.S. (2010). Use of associationrules to explore cause�effect relationships in occupationalaccidents in the Taiwan construction industry. SafetyScience, 48(4), 436�444.
Courtney, T.K., Sorock, G.S., Manning, D.P., Collins, J.W., &Holbein-Jenny, M.A. (2001). Occupational slip, trip, andfall-related injuries�can the contribution of slipperiness beisolated? Ergonomics, 44(13), 1118�1137.
Dudarev, A.A., Karnachev, I.P., & Odland, ;.J. (2013). Occupa-tional accidents in Russia and the Russian Arctic. Interna-tional Journal of Circumpolar Health, 72, 32�32.
Dumrak, J., Mostafa, S., Kamardeen, I., & Rameezdeen, R.(2013). Factors associated with the severity of constructionaccidents: The case of South Australia. Australasian Journalof Construction Economics and Building, 13(4), 32�49.
Giudici, P. (2003). Applied data mining: Statistical methods forbusiness and industry. New York, NY: Wiley.
Halvani, G.H., Jafarinodoushan, R., Mirmohammadi, S.J., &Mehrparvar, A.H. (2012). A survey on occupational acci-dents among construction industry workers in Yazd city:Applying time series 2006�2011. Journal of OccupationalHealth and Epidemiology, 1(1), 1�8.
Han, J., & Kamber, M. (2001). Data mining: Concepts and tech-niques. China Machine Press, 8, 3�6.
Huang, X., & Hinze, J. (2003). Analysis of construction workerfall accidents. Journal of Construction Engineering andManagement, 129(3), 262�271.
Im, H.J., Kwon, Y.J., Kim, S.G., Kim, Y.K., Ju, Y.S., & Lee, H.P. (2009). The characteristics of fatal occupational injuriesin Korea’s construction industry, 1997�2004. Safety Sci-ence, 47(8), 1159�1162.
International Labour Organization (2013). Retrieved from http://www.ilo.org/safework/events/meetings/WCMS_204594/lang�en/index.htm.
Ivancic, P.C. (2013). Hybrid cadaveric/surrogate model of thora-columbar spine injury due to simulated fall from height.Accident Analysis & Prevention, 59, 185�191.
Liao, C.W., & Perng, Y.H. (2008). Data mining for occupationalinjuries in the Taiwan construction industry. Safety Science,46(7), 1091�1102.
Liao, C.W., Perng, Y.H., & Chiang, T.L. (2009). Discovery ofunapparent association rules based on extracted probability.Decision Support Systems, 47(4), 354�363.
Lin, Y.H., Chen, C.Y., & Wang, T.W. (2011). Fatal occupationalfalls in the Taiwan construction industry. Journal of the Chi-nese Institute of Industrial Engineers, 28(8), 586�596.
Ling, F.Y.Y., Liu, M., & Woo, Y.C. (2009). Construction fatali-ties in Singapore. International Journal of Project Manage-ment, 27(7), 717�726.
L�opez Arquillos, A., Rubio Romero, J.C., & Gibb, A. (2012).Analysis of construction accidents in Spain, 2003�2008.Journal of Safety Research, 43(5), 381�388.
Loudoun, R.J. (2010). Injuries sustained by young males in con-struction during day and night work. Construction Manage-ment and Economics, 28(12), 1313�1320.
Mingers, J. (1989). An empirical comparison of pruning methodsfor decision tree induction.Machine Learning, 4(2), 227�243.
Nenonen, N. (2012). Analysing factors related to slipping, stum-bling, and falling accidents at work: Application of datamining methods to Finnish occupational accidents and dis-eases statistics database. Applied Ergonomics, 44, 215�224.
Parhizi, S., Shahrabi, J., & Pariazar, M. (2009). A new accidentinvestigation approach based on data mining techniques.Journal of Applied Sciences, 9(4), 731�737.
P�erez-Alonso, J., Carre~no-Ortega, �A., V�azquez-Cabrera, F.J., &Callej�on-Ferre, �A.J. (2012). Accidents in the greenhouse-construction industry of SE Spain. Applied Ergonomics, 43(1), 69�80.
Persona, A., Battini, D., Faccio, M., Bevilacqua, M., & Ciarap-ica, F.E. (2006). Classification of occupational injury casesusing the regression tree approach. International Journal ofReliability, Quality and Safety Engineering, 13(02),171�191.
Salminen, S. (2004). Have young workers more injuries thanolder ones? An international literature review. Journal ofSafety Research, 35(5), 513�521.
Scallan, E., Staines, A., Fitzpatrick, P., Laffoy, M., & Kelly, A.(2001). Injury in Ireland (Report No. 15042) (Report of theDepartment of Public Health Medicine and Epidemiology).Dublin: University College Dublin.
Social Security Organization of the Islamic Republic of Iran.(2012). Statistical report of occupational accidents in theconstruction industry between 2007�2011. Tehran: Statis-tics and Social-economic Calculations Office.
Su�arez-Cebador, M., Rubio-Romero, J.C., & L�opez-Arquillos,A. (2014). Severity of electrical accidents in the constructionindustry in Spain. Journal of Safety Research, 48, 63�70.
Tam, C.M., Zeng, S.X., & Deng, Z.M. (2004). Identifying ele-ments of poor construction safety management in China.Safety Science, 42(7), 569�586.
Wang, H.S., Yeh, W.C., Huang, P.C., & Chang, W.W. (2009).Using association rules and particle swarm optimizationapproach for part change. Expert Systems with Applications,36(4), 8178�8184.
Warch, S.L. (2002). Quantifying the financial impact of occupa-tional injuries and illnesses, and the costs and benefits asso-ciated with an ergonomic risk control intervention withinthe unapprised business segment of UnitedHealth group(Unpublished doctoral dissertation). Menomonie, WI: Uni-versity of Wisconsin-Stout.
Wojtczak-Jaroszowa, J., & Jarosz, D. (1987). Time-related dis-tribution of occupational accidents. Journal of SafetyResearch, 18(1), 33�41.
International Journal of Injury Control and Safety Promotion 13