Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) prediction model based on limited peat...

Area of Science:

• Environmental Chemistry
• Computational Toxicology
• Machine Learning Applications

Background:

• Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) are highly toxic by-products of incomplete combustion.
• Malaysian peat soils, often acidic with high chlorine content, can promote PCDD/Fs formation.
• Accurate prediction of PCDD/Fs emissions is crucial for environmental protection and public health.

Purpose of the Study:

• To develop and optimize an artificial neural network (ANN) model for predicting PCDD/Fs emissions in peat soil.
• To improve ANN prediction accuracy by optimizing initial connection weights and bias using differential evolution (DE).
• To assess the viability of an evolutionary-optimized ANN methodology for environmental monitoring of PCDD/Fs.

Main Methods:

• Utilized a multilayer perceptron (MLP) with a backpropagation algorithm for PCDD/Fs prediction.
• Employed differential evolution (DE), specifically self-adaptive ensemble-based differential evolution with enhanced population sizing (SAEDE-EP), to optimize ANN parameters.
• Trained and tested various ANN architectures using real-world datasets with eight input variables and one output variable.

Main Results:

• The optimized ANN, featuring 5 hidden neurons and SAEDE-EP, achieved the lowest test mean squared error (MSE_test) of 6.1790 × 10⁻³.
• The model demonstrated high predictive accuracy, evidenced by the highest R² value of 0.97447.
• The evolutionary-optimized ANN methodology proved effective in predicting PCDD/Fs emissions.

Conclusions:

• An evolutionary-optimized ANN-based approach is a cost-effective and viable solution for predicting PCDD/Fs emissions in peat soil.
• This methodology can significantly aid pollution control, environmental monitoring, and authorities in preventing human exposure to PCDD/Fs.
• The study highlights the potential of advanced computational methods in addressing environmental contamination issues.