Enhanced Predictive Survival Model for Bone Marrow Transplantation Using Optimized Machine Learning Algorithms

Main Article Content

Pawan Bishnoi
Shailesh Kumar
Rohit Bajaj

Abstract

Bone marrow transplantation (BMT) is an important medical technique for many hematological illnesses; yet, post-transplant survival rates are extremely varied due to complicated clinical and biological variables. Accurate prediction of patient survival is critical for making better clinical decisions and risk stratification. This work proposes an improved predicted survival model for bone marrow transplantation that employs optimized machine learning techniques. A complete data-driven framework is created to include patient demographics, clinical characteristics, and transplant-related factors. To increase prediction performance, many machine learning models are applied and systematically optimized utilizing feature selection and hyper parameter tuning approaches. These models include Decision Tree, Random Forest, Support Vector Machine, and Neural Network. Standard performance indicators, including accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC), are used to evaluate the model. This improved model outperforms baseline models in terms of predicted accuracy and resilience, demonstrating the efficacy of optimisation procedures when dealing with high-dimensional clinical data. The findings show that the suggested technique may be used as a viable clinical decision support tool to predict post-transplant survival outcomes. This study advances intelligent healthcare systems by allowing for early risk assessment and personalised treatment planning for individuals undergoing bone marrow transplantation.

Article Details

Section
Articles