THE PREDICTION AND INFLUENCING FACTORS OF BREAST CANCER RECURRENCE BASED ON RANDOM FOREST

Breast cancer (BC) is one of the most common malignant tumors in women. In 2022, it has become the second most common cancer after lung cancer. Although medical technology has made great progress in recent years and the survival rate of breast cancer patients has been greatly improved, according to research, about 40% of patients still relapse after treatment. Constructing a breast cancer recurrence prediction model and finding factors that affect breast cancer recurrence are of great significance for clinical treatment and prolonging patient survival. This study used the TCGA dataset and randomly divided the patients into training and test sets in a ratio of 8:2. Seven algorithms, including decision tree, logistic regression, support vector machine, K nearest neighbor, random forest, neural network, and adaptive boosting, were used to construct the model, and the performance of each model was evaluated. The results showed that the random forest model had the best effect, with an accuracy of 97.77%, a sensitivity of 94.23%, a specificity of 99.21%, a false positive rate of 0.79%, an F1 of 96.08%, and an AUC value of 96.72%. The features obtained by the model classification were ranked according to their importance. The top three features were: Age_at_Initial_Pathologic_Diagnosis_nature2012, lymph_node_examined_count and number_of_lymphnodes_Postive _by_he. The model provides more robust feature importance analysis results, providing an important reference for clinicians in breast cancer recurrence risk assessment and individualized treatment decision-making.