Bayesian algorithm implementation in a real time exposure assessment model on benzene with calculation of associated...

Area of Science:

• Occupational Health and Safety
• Environmental Science
• Computational Toxicology

Background:

• Gas station employees face potential health risks from benzene exposure.
• Accurate real-time exposure assessment is crucial for effective risk management.
• Existing models may not fully capture dynamic environmental factors influencing exposure.

Purpose of the Study:

• To develop a reliable modeling platform for real-time personal exposure and health risk calculation for gas station employees.
• To predict benzene exposure patterns using Artificial Neural Networks (ANNs).
• To assess leukemia risk using a Physiology Based Pharmaco-Kinetic (PBPK) model integrated with ANN data.

Main Methods:

• Development of an Artificial Neural Network (ANN) model using Bayesian regularization to predict benzene exposure.
• Construction of a Physiology Based Pharmaco-Kinetic (PBPK) risk assessment model.
• Integration of ANN predictions into the PBPK model for leukemia risk calculation.
• Utilization of Bayesian algorithms and Monte Carlo simulations for risk assessment.

Main Results:

• Bayesian regularization demonstrated superior performance for the ANN exposure model.
• The integrated modeling system effectively estimated personal exposure and associated health risks in real-time.
• The system was successfully evaluated in both urban and rural filling station environments.

Conclusions:

• The developed modeling platform provides a reliable method for real-time occupational health risk assessment for gas station employees.
• The system effectively utilizes environmental sensor data to estimate benzene exposure and leukemia risk.
• This approach offers a promising tool for proactive occupational health management in similar environments.