 The study aimed to predict the risk of kidney graft failure across three temporal cohorts using machine learning-based classification algorithms and deep learning-based auto-encoders for data dimensionality reduction. The results showed that the models predicted graft survival with high accuracy and feature importance analysis revealed varying influences of clinical features on graft survival across different time periods. This article was authored by Syed Assel Ali Naqvi, Karthik Tenenkor, Amanda Vincent, and others.