Sobhan Sarkar

Text mining based safety risk assessment and prediction of occupational accidents in a steel plant

Occupational accidents are a serious threat to any organization. Occupational accidents in steel industry sector remain a threat as workforce is exposed to different kinds of hazards due to the workplace characteristics. In this study, a unique method is proposed by developing a text mining based prediction model using fault tree analysis (FTA), and Bayesian Network (BN). Free unstructured accident dataset for a period of four years has been used in this study. Text mining approach results in finding the basic events concerning each of primary causes. The basic events, in turn, are utilized in building FT and BN diagram that could predict the occurrence of accidents attributable to different primary causes. The model, so developed, can be considered adequate with 87.5% accuracy. Furthermore, sensitivity analysis is performed for the validation of the model.

Study of optimized SVM for incident prediction of a steel plant in India

Occupational accident is a serious issue for every industry. Steel industry is considered to be one of the economic sectors having a high number of accidents. Thus, the main aim of this study is to build a model which could predict the occupational incidents (i.e., injury, near-miss, and property damage) using support vector machine (SVM) by utilizing a database comprising almost 5000 occupational accidents reports from an integrated steel plant corresponding to the span of years 2010 to 2012. Parameter optimization of the SVM is performed using grid search (GS), genetic algorithm (GA), and BAT algorithm to obtain the better accuracy of the classifier. The results of experiments show that grid search-based SVM outperforms other optimized SVM approaches with 88.0% accuracy. Other optimization techniques can also be adapted to search for the better prediction accuracy of the model.

Prediction of occupational accidents using decision tree approach

The focus of the present study is to build a predictive model which not only could predict the occupational incidents but also provide rules for explaining accident scenarios like near-miss, property damage, or injury cases. Classification and regression tree (CART) is used for prediction purpose. Furthermore, the parameters of CART have been tuned by grid based tuning and genetic algorithm (GA). The experimental results show that the GA optimized CART provides better accuracy than others. Additionally, the best rules extracted from GA optimized CART are discussed in order to adopt better safety precautionary measures at work.

Supplier Selection in Uncertain Environment: A Fuzzy MCDM Approach

This paper addresses a critical issue of selection of supplier occurred in supply chain of a manufacturing company. As there are lot more criteria present for decision making of suitable supplier selection among many, it becomes more challenging task for any company to make as this decision is entangled with company’s profit and time. So, to address this problem, this paper proposes a multi-criteria decision making (MCDM) method using Decision Making Trial and Evaluation Laboratory (DEMATEL) based on Analytic Network Process (ANP), i.e., DANP, with fuzzy Vise Kriterijumska Optimizacija I Kompromisno Resenje (FVIKOR) to judiciously select suppliers based on important criteria and to point out interrelationships among dimensions and criteria in SCM by Network Relationship Map (NRM) for this company. Furthermore, the ranking is supported by sensitivity analysis.

Measurement and Modeling of Job Stress of Electric Overhead Traveling Crane Operators

Background In this study, the measurement of job stress of electric overhead traveling crane operators and quantification of the effects of operator and workplace characteristics on job stress were assessed. Methods Job stress was measured on five subscales: employee empowerment, role overload, role ambiguity, rule violation, and job hazard. The characteristics of the operators that were studied were age, experience, body weight, and body height. The workplace characteristics considered were hours of exposure, cabin type, cabin feature, and crane height. The proposed methodology included administration of a questionnaire survey to 76 electric overhead traveling crane operators followed by analysis using analysis of variance and a classification and regression tree. Results The key findings were: (1) the five subscales can be used to measure job stress; (2) employee empowerment was the most significant factor followed by the role overload; (3) workplace characteristics contributed more towards job stress than operators characteristics; and (4) of the workplace characteristics, crane height was the major contributor. Conclusion The issues related to crane height and cabin feature can be fixed by providing engineering or foolproof solutions than relying on interventions related to the demographic factors.

Data-driven Mapping Between Proactive and Reactive Measures of Occupational Safety Performance

This study aims to analyse the incident investigation reports logged after the occurrence of events from an integrated steel plant and map it with proactive safety data. From the narrative text describing the event, this study has attempted to unfold the hazards and safety factors present at the workplace. Text document clustering with expectation maximization algorithm (EM) has been used to group the different events and find key phrases from them. These key phrases are considered as the root causes of the reported events. This study shows how the mapping of the safety factors from both proactive safety data and incident reports can help in the improvement of safety performance as well as better allocation of resources. The study points out specific areas to the management where improvements are needed. The mapping also indicates the areas of improvement made by the constant effort of safety practitioners.

Prediction of Occupational Incidents Using Proactive and Reactive Data: A Data Mining Approach

Prediction of occupational incidents is an important task for any industry. To do this, reactive data has been used by most of the previous studies in this domain. As an extension of the existing works, the present study has used the underused proactive data coupled with reactive data to establish the predictive models so that the information inherent in both data sets could be better utilized. The main aim of the study is to predict the incident outcomes using mixed data set comprising reactive and proactive data together. Two decision tree classifiers, i.e. classification and regression tree (CART) and C5.0, have been implemented with tenfold cross validation. Furthermore, the ensemble technique, namely adaptive boosting has been implemented to increase the classification accuracy. Results show that boosted C5.0 produces higher accuracy than others for the prediction task. Furthermore, the rules obtained produce the insight of the incidents. The limitation of the present study includes the use of less amount of data and the requirement of experts’ domain knowledge for a large span of time. Future scope of the study includes the proper feature selection for preparation of the mixed data set and building the better classification algorithm for better prediction of occurrence of accidents. The present work sets out the potential use of both types of data sources together.

Genetic Algorithm-Based Association Rule Mining Approach Towards Rule Generation of Occupational Accidents

Occupational accident is a grave issue for any industry. Therefore, proper analysis of accident data should be carried out to find out the accident patterns so that precautionary measures could be undertaken beforehand. Association rule mining (ARM) technique is mostly used in this scenario to find out the association (i.e., rules) causing accidents. But, among the rules generated by ARM, all are not useful. To handle this kind of problem, a new model ARM and genetic algorithm (GA) has been proposed in this study. The model automatically selects the optimal Support and Confidence value to generate useful rules. Out of 1285 data obtained from a steel industry in India, eleven useful rules are generated using this proposed method. The findings from this study have the potential to help the management take the better decisions to mitigate the occurrence of accidents.

Oil Spill Detection Using Image Processing Technique: An Occupational Safety Perspective of a Steel Plant

Oil spill at workplace is one of the potential hazards in industry. Though it has not attracted more importance from research point of view, it can lead to economic loss for the industry through the occurrence of accident phenomena like slipping, firing, or pollution to the environment. Hence, oil spill detection should be considered as an essential research issue. In order to address this, the present study endeavors to use image processing technique for oil spill detection using the image data retrieved from an integrated steel plant in India. Results reveal that the technique adopted for oil spill detection is an effective and efficient way. This method, though used in steel plant, can be used in any other industry like construction, manufacturing.

Decision Support System for Prediction of Occupational Accident: A Case Study from a Steel Plant

Decision support system (DSS) is a powerful tool which helps decision-makers take unbiased and insightful decisions from the historical data. In the domain of occupational accident analysis, decision-making should be effective, insightful, unbiased, and more importantly prompt. In order to obtain such decision, development of DSS is necessary. In the present study, an attempt has been made to build such DSS for accident analysis in an integrated steel plant. Two classifiers, i.e., support vector machine (SVM) and random forest (RF) have been used. RF produces better level of accuracy, i.e., \(99.34\%\). The developed DSS has full potential in making insightful decisions and can be used in other domains like manufacturing, construction, etc.