Abstract
Background: Diabetes affects cardiovascular diseases, vision loss, and kidney disease. Data mining techniques help doctors predict the disease early and treat it accurately. Given the role of physical activity in preventing diabetes, this study examined its effect alongside other risk factors. Therefore, the main objective of this study was to predict diabetes in active men using the Adaboost algorithm.
Methods: In this applied-developmental study, the records of 500 male patients aged 20-80 years with a history of regular physical activity (at least 3 90-minute sessions per week) who had visited Milad and Ayatollah Kashani hospitals in Tehran in the past 10 years were selected as samples. 20 anthropometric, genetic, family history, environmental and physiological variables were selected as input features of the algorithm. The criteria of sensitivity, accuracy and precision were used to evaluate the performance of the proposed algorithm. MATLAB (2024) software was used for data analysis.
Results: The results showed that the Adaboost algorithm was able to predict diabetes in active men with an accuracy of 73.1% and precision of 75.3%.
Conclusion: This model can be used as an auxiliary tool in the initial screening of diabetes in medical centers.