Occurrence Type Classification for Establishing Prevention Plans Based on Industrial Accident Cases Using the KoBERT Model
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors present research related to the occurrence and type classification for establishing prevention plans based on industrial accident cases using the KoBERT model.
The motivation of this research includes, among other things, the statement of increasing industrial sophistication and complexity leading to workplaces to become increasingly prone to industrial accidents, and to the fact that South Korea has implemented new policies addressing potential risks to overcome stagnation in industrial accident reduction and predict site accidents from past cases.
The authors state that human errors, subjective judgments, synonyms, and terms incorrectly used by classifiers reduce original data quality and impede developments or applications of policies, technologies, and methods preventing accidents based on past accidents.
To overcome this shortcoming the authors propose three artificial intelligence models to objectively classify the occurrence type of accident cases. Models are developed based on a natural language processing model (KoBERT), which considers Korean language characteristics.
Basically, the authors carried out an adequate work in elaborating AI based models within the initial scope of their research. However, there is a basic issue related to root causes of accidents that are not captured in the model/scope of research. The findings from the research works could not help reversing stagnation in industrial accident reduction and predict site accidents from past cases. It is discussed below.
The proposed scope/models seem too narrow not going deep enough in the analysis given that they do not adequately take into account the complexity of the operational environment (there are new hazards and risks related to the complexity of operational environment and organizational and human performance) as well as the influence of biases in decision making (motivational and cognitive) leading to accidents. Safety culture/management role are not adequately taken into consideration and the models do not capture it.
Such an approach often leads to putting the blame on frontline workers for deficiencies which are at the organizational level enabling and tolerating conditions creating unsafe workplaces (leading to the “drift to failure/accident”). The practice shows that these influences are “soft factors” that are hard to resolve, and may easily be hidden by apparent factors. Recent research works and return of experience show that the organizational performance and safety culture play a key role in creating conditions for accidents. The organizational performance also includes less studied motivational biases in decision-making process (mentioned above), and it is not considered either (cognitive biases are relatively well studied and understood).
Thus, the scope of the paper does not enable to capture the true image of the analyzed topic. Given the limited scope of the paper, the literature review is also too narrow, and it should be expanded to include aspects discussed above.
It is strongly recommended that the authors consult the following references while revising the paper.
Brocal, F., González-Gaya, C., Komljenovic, D., Katina, P.D., Sebastián, M.A., (2019), Emerging risk management in Industry 4.0: an approach to improve organizational and human performance in the complex systems, Complexity, Vol., 2019, Article ID 2089763, https://doi.org/10.1155/2019/2089763
Dekker S, Cilliers P, Hofmeyr, J.H., (2011), The complexity of failure: Implications of complexity theory for safety investigations. Safety Science. 49: 939-945
Kahneman, D. Thinking, Fast and Slow, Farrar, Straus and Giroux, New York: 2012
Komljenovic, D., Loiselle, G., Kumral, M., (2017), Organization: a new focus on mine safety improvement in a complex operational and business environment, International Journal of Mining Science and Technology, 27: 617-625
Leveson N.G., (2011a), Engineering a Safer World, Systems Thinking Applied to Safety. Cambridge MA: The MIT Press
Leveson N.G., (2011b), Applying system thinking to analyze and learn from events. Safety Science, 49:55-64
Montibeller, G. and Winterfeldt, D. Cognitive and Motivational Biases in Decision and Risk Analysis, Risk Anal. 2015: 35 (7): 1230-1251
Mosey, D. Looking beyond operator – Putting people in the mix, NEI Magazine, 2014; http://www.neimagazine.com/features/featurelooking-beyond-the-operator-4447549/
in collaboration with Ken Ellis, Managing Direction of World Association of Nuclear Power Operators (WANO)
http://www.neimagazine.com/features/featureputting-people-in-the-mix-4321534/
http://www.neimagazine.com/features/featureputting-people-in-the-mix-part-2-4322674/
Stacey, R.D. and Mowles, C. (2016) Strategic Management and Organisational Dynamic; The Challenge of Complexity to Ways of Thinking about Organisations (seventh edition), Pearson, London, New York, Boston, Paris.
Author Response
Thank you very much for your detailed comments and suggestions. We have revised the paper based on the comments, as shown in the attached file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsI read with interest the article: Occurrence Type Classification for Establishing Prevention Plans based on Industrial Accident Cases using the KoBERT model. However, I have comments that the authors should incorporate:
1. I recommend changing the title and editing, abstracting and adding key words so that they are in line with the content of the article, as in the current form it seems as if the authors want to focus on industrial accidents (disaster management) and not on Occupational accidents and the issue of Occupational Safety and Health. Also, preventive measures themselves are not given the necessary attention in the article.
2.The Indroduction is relatively general and focuses primarily on safety and health protection. The issue of crisis management and risk management should be supplemented, as well as the impact of these accidents on the company and society. Also, the importance of table 1 is not clearly illustrated in the text. The introduction should end with a clearly defined goal of the article.
3. I recommend the authors to standardly divide the article into other chapters for the sake of clarity (Introduction, Materials and Methods, Results, Discussion, and Conclusions). The article lacks a clear discussion that would clarify the whole issue. The benefits of the article as well as its limitations should be presented more clearly in the discussion. The individual claims in this section need to be supported by a number of other research sources that have dealt with similar issues.
Author Response
We have revised the paper based on the comments, as shown in the attached file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors adequately addressed the previous comments in the revised paper.
Author Response
Thank you very much for your detailed comments and suggestions.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors mostly incorporated my comments. But it is necessary, they paid even more attention in the Discussion and Future Research section. In this part, I recommend that individual paragraphs be more detailed and better illustrate the potential of the entire article and further studies following the article.
Author Response
Thank you very much for your detailed comments and suggestions. The attached file is the summary of the responses and revisions. We believe the manuscript has been improved, and hope it will be accepted for publication.
Author Response File: Author Response.pdf
