Machine Learning Applications in Personalised Medicine: Predictive Models for Disease Risk Assessment and Treatment Optimisation in Biomedical Data Analysis

This study investigates the use of machine learning (ML) techniques in personalised medicine, specifically focusing on the development of predictive models for disease risk assessment and treatment optimisation. By utilising various data sets, including genomic, clinical, and imaging data, four ML algorithms—Random Forest, Support Vector Machine, Gradient Boosting, and Neural Network—were evaluated for their effectiveness in generating accurate predictions. The results show that Gradient Boosting demonstrated superior performance across all data sets, achieving accuracies of 0.88, 0.85, and 0.83, and areas under the ROC curve (AUC) of 0.92, 0.90, and 0.88 on genomic, clinical, and imaging data, respectively. Neural Network also showed competitive performance with accuracies of 0.87, 0.84, and 0.81, and AUCs of 0.91, 0.89, and 0.87 on the same data sets. Random Forest and Support Vector Machine achieved accuracies ranging from 0.80 to 0.85 and AUCs from 0.86 to 0.90. These findings highlight the potential of ML algorithms in leveraging diverse biomedical data to develop robust predictive models, thereby facilitating personalised healthcare interventions. The study contributes to advancing the understanding and implementation of ML in personalised medicine, with recommendations for improving clinical decision-making and patient outcomes.