'MONITORING AND EVALUATION OF RADON GAS POTENTIAL USING MACHINE LEARNING ALGORITHMS'

This master's project focuses on developing a robust methodology for assessing indoor radon gas potential using machine learning algorithms. Radon, an odorless and colorless radioactive gas, poses significant health risks, including lung cancer. Addressing this public health concern, the study aims to create a reliable prediagnosis model and an accessible web platform for evaluating radon potential in buildings. The project leverages a comprehensive dataset, gathered from various building types and conditions, to train and test machine learning models. Decision trees and random forests were selected for their interpretability and accuracy. Despite the promising methodology, the model achieved a 70% accuracy rate, primarily due to the limited quantity of training data, necessitating careful consideration of the estimates provided. The developed web platform integrates this machine learning model, allowing users to input building characteristics and receive immediate radon potential assessments. This platform not only facilitates individual awareness and risk mitigation but also serves as a tool for collecting further data to enhance the model's accuracy over time. Key features include user-friendly input forms, data visualization, and an administrative dashboard for monitoring performance and engagement. Significantly, the project's outcomes have been disseminated through publications and presentations at prominent conferences. The papers, "A Pre-Diagnosis Model for Indoor Radon Potential Assessment" and "Citizen Science Platform for Radon Potential Assessment and Radon Risk Awareness", highlight the model's potential and the platform's efficacy in raising public awareness and contributing to public health initiatives. In conclusion, this research advances the field of environmental health by providing a practical tool for radon risk assessment, with the potential for widespread application and continuous improvement through citizen science contributions. ; A investigação foi financiada pelo projeto TECH ( ...