Abstract

Bayesian approach is a quantitative risk analysis technique allowing a probabilistic analysis of dreaded event with their causes and effects. By creating a Bayesian Belief etwork a model knowledge can be developed to predict the studied system evolution. The aim of this method is to estimate a posteriori probability of the dreaded event, their causes and consequences, and gives a developed decisional support tool for risk assessment.In this paper, Bayesian approach will be experimented on safety data of a cement factory in west of Algeria (CFWA). Accident data of 6 years are saved, however some records are not completed or description is not clear generating incompleteness and uncertainty. A conceptual and generic model should be developed with a protocol for assessing the risk of injuries on persons and a computational module which constitute a decisional support system for risk assessment (DSRS). Then, experiment the model and generate many scenarios to understand and create knowledge about causes or conditions that generate those dreaded events. And finally, based on the results of this research, problem areas were determined such as struck by falling objects is a major hazard at cement industry, and convince the staff not to neglect the instructions to wear individual protection equipment (IPE).