Multinomial classification to predict the most effective adjuvant combination therapies for breast cancer patients

Abstract

Accurately predicting effective treatment methods based on personalized tumor genetic profiles is a major goal of precision cancer medicine. Because people with breast cancer at comparable stages respond differently to treatment, it is essential to gain insight into the variables that influence treatment success. This study presents a supervised multinomial logistic regression model for predicting the best adjuvant therapy for breast cancer patients to lower the probability of metastatic recurrence. This model will assist health professionals (physicians) in making judgments about which medicinal regimens to suggest to patients. In addition, this article presents a comparison of several multinomial machine learning methods (Logistic Regression (LR), Naive Bayes (NB), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), and Neural Network (ANN)). The results reveal that the Random Forest classifier is more effective in terms of adjuvant therapy combination prediction accuracy.