Utilization of Machine Learning Models for Forecasting Survival in Cancer Patients: A Comparative Study Employing Logistic Regression, AdaBoost, SVM, and Random Forest

This research explores the utilization of machine learning algorithms to forecast patient survival outcomes utilizing a synthetic cancer dataset. A thorough data preprocessing pipeline, which includes sophisticated feature engineering and selection methods, was implemented to enhance data quality and relevance. Four classification algorithms were assessed, with ensemble techniques such as AdaBoost and Random Forest significantly surpassing conventional methods due to their capacity to model intricate, non-linear interactions. Stringent hyperparameter optimization and cross-validation were conducted to ensure the models' robustness and generalizability, addressing potential overfitting and improving predictive reliability. The results highlight the essential importance of data preprocessing, model selection, and systematic evaluation in creating effective machine learning solutions for healthcare. This research illustrates the transformative potential of machine learning in predictive analytics, providing valuable insights into its application for personalized medicine and evidence-based clinical decision-making. By demonstrating the synergy between advanced algorithms and high-quality data, this study aids in the progression of machine learning applications within the healthcare sector.

کلیدواژه ها:

Machine Learning ، Cancer Survival Prediction ، Feature Importance ، Ensemble Learning ، models comparison