Abstract Breast cancer is one of the most common types of cancer among women, an‎d its early detection an‎d prevention can play a vital role in reducing mortality an‎d treatment costs. The aim of this study was to identify the factors influencing breast cancer prevention using data mining algorithms. To this end, data were collected through a questionnaire from two groups of women: healthy individuals an‎d patients diagnosed with breast cancer. The dataset, which included individual characteristics, lifestyle factors, an‎d medical history, was augmented through simulation an‎d then divided into three subsets: training (735 samples), validation (157 samples), an‎d testing (158 samples). For analysis an‎d modeling, three machine learning algorithms were applied: Decision Tree, Ran‎dom Forest, an‎d Support Vector Machine. The eva‎luation results on the test set revealed that the Decision Tree algorithm achieved the best performance with an accuracy of 89.2% an‎d an F1-score of 0.89, followed by Ran‎dom Forest with an accuracy of 83.5% an‎d an F1-score of 0.83. In contrast, Support Vector Machine showed weaker performance with an accuracy of 64.5% an‎d an F1-score of 0.55. Confusion matrices an‎d classification reports confirmed these differences. Feature importance analysis indicated that variables such as menopause status, age at menopause, weight, breastfeeding, stress, smoking, an‎d alcohol consumption were among the most significant predictors of breast cancer. The findings suggest that a healthy lifestyle, weight control, an‎d educational an‎d healthcare interventions can play a crucial role in preventing the disease. Ultimately, data mining models—particularly interpretable algorithms such as Decision Trees—can serve as valuable tools for medical data analysis an‎d the development of preventive strategies.

علي اصغر بازدار

معصومه مسي بيدگلي