Data mining for pesticide decontamination using heterogeneous photocatalytic processes

Area of Science:

• Environmental Chemistry
• Computational Chemistry
• Artificial Intelligence

Background:

• Pesticides are essential for agriculture but pose environmental and health risks.
• Heterogeneous photocatalysis is a promising method for pesticide degradation.
• Developing predictive models for photocatalytic decontamination is crucial for environmental safety.

Purpose of the Study:

• To apply artificial intelligence (AI) techniques for optimizing pesticide decontamination using heterogeneous photocatalysis.
• To develop and compare predictive models for pesticide removal efficiency.

Main Methods:

• A systematic literature review of 537 cases from 45 articles (2000-2020).
• Application of Cross-Industry Standard Process (CRISP) methodology.
• Evaluation of four machine learning classifiers: Decision Trees (DT), Bayesian Network (BN), Support Vector Machines (SVM), and Feed Forward Multilayer Perceptron Neural Networks (MLP).
• Performance assessment using accuracy and sensitivity metrics.

Main Results:

• The Decision Trees (DT) model, utilizing seven predictive factors, demonstrated superior performance.
• DT achieved an accuracy of 91.06% and a sensitivity of 80.32%.
• This indicates DT's effectiveness in predicting pesticide decontamination outcomes.

Conclusions:

• AI-driven predictive models, particularly Decision Trees, can significantly enhance the efficiency of pesticide decontamination processes.
• The developed DT model offers a reliable tool for optimizing heterogeneous photocatalytic degradation of pesticides.
• This research contributes to safer environmental practices through advanced computational methods.