An Exploratory Study Using an Artificial Neural Network to Predict Fatigue from Mobile Phone Use: A Population-Based Survey in Algeria
International Journal of Medical Toxicology and Forensic Medicine,
Vol. 16 (2026),
1 January 2026,
Page 1-9
https://doi.org/10.22037/ijmtfm.v16.50757
Background: Mobile phones are widely used, raising questions about the possible health effects of the radiofrequency electromagnetic fields they emit. Fatigue is frequently reported in this context, but it is influenced by multiple, interacting factors that are difficult to model with classical statistical methods. Artificial neural networks (ANNs) may help explore these complex relationships.
Methods: An ANN model was developed to estimate fatigue associated with cell phone use. Data were collected via a semi-structured questionnaire completed by 478 Algerian participants. The survey recorded sociodemographic data and patterns of mobile phone use. The network had a 5-10-1 architecture, and its performance was evaluated by mean squared error (MSE) and the coefficient of determination (R²). A simple MATLAB interface was created to allow user input and display model outputs with a colour-coded indicator.
Results: The ANN achieved an MSE of 0.5993, indicating that it reproduced some general patterns in the data. However, the coefficient of determination was low (R² = 0.0338), showing that only a small proportion of the variability in fatigue scores was explained and that individual predictions were imprecise.
Conclusion: This exploratory study suggests that ANN-based models are feasible for analysing fatigue associated with mobile phone use, but the findings should be regarded as preliminary and are subject to some limitations. Larger, more diverse samples and richer, preferably objective, exposure and health measures will be required before such tools can be used for reliable risk assessment or public health guidance.