In strategic planning of oil spill, it is crucial to understand what could be the spill size given incident properties, and most current engineering models rely on physical simulation to get spill size of individual spill incidents.
This project aims to predict spill size based on oil spill incidents, and since it is using machine learning based models, it is faster than its counterparts.
to predict damage and oil outflow in tanker collision accidents
(i) Deep Neural Network, (ii) Gradient Boosted Regression Tree and (iii) Polynomial Regression based models are developed and trained on simulated accident data based on Monte Carlo Simulation
The proposed DNN is highly accurate and computationally efficient.
