Idiopathic pulmonary fibrosis (IPF) ranks among the most frequent forms of interstitial lung disease and is characterized by progressive scarring (fibrosis) of lung tissue. Affected individuals are generally advised to pursue lung transplantation; failure to do so often leads to ongoing, irreversible pulmonary deterioration and eventual mortality. With advanced, irreversible damage, reliable forecasting of patient survival becomes critically important. Conventional clinical approaches typically lack robust predictive instruments for this purpose. Nevertheless, the proven capacity of artificial intelligence to address complex medical scenarios has opened the door to mortality prediction through machine learning methodologies. The present study introduced a soft voting ensemble classifier based on the 30 most informative clinical features to estimate mortality risk in patients diagnosed with idiopathic pulmonary fibrosis. Five distinct machine learning models were integrated for this task: random forest (RF), support vector machine (SVM), gradient boosting machine (GBM), XGBoost (XGB), and multi-layer perceptron (MLP). Integration of the classifiers via the soft voting ensemble yielded an accuracy of 79.58%, sensitivity of 86%, F1-score of 84%, prediction error of 0.19, and responsiveness of 0.47. Conclusions: The developed model can assist physicians in clinical decision-making, longitudinal disease monitoring, and risk mitigation. Ultimately, it supports efforts to lower mortality rates, optimize patient health status, and improve risk stratification.
