A Machine Learning-Based Predictive Model for Drug Sensitivity in Breast Cancer Using Gene Expression Data

Abstract

Through the combination of tool learning patterns, this study offers a novel strategy for personalised treatment for the majority of breast malignancies. The authors used a carefully assembled dataset that included 3444 cases of drug management data, affected person profiles, diagnostic scans, and scientific reviews to train artificial neural networks (ANN), support vector machines (SVM), decision trees (DT), and random forests (RF) for drug sensitivity prediction modelling. While SVM demonstrated its capacity to handle high-dimensional statistics with an accuracy of 96.5%, the artificial neural network (ANN) exhibited remarkable versatility, achieving a commendable accuracy rate of 97.5%. The interpretability inherent in decision trees (DT) and the combined energy of random forests (RF) added crucial elements to the multifaceted methodology. The outcome of the research underscores that the proposed machine learning model stands out with the highest efficacy in predicting the most accurate drug for a given patient.