Liver transplantation (LT) continues to increase as a potentially curative intervention for patients with end-stage liver disease, yet success hinges on a delicate balance of donor-recipient match and accurate survival prediction. The complexity of the task is then addressed in this study using a novel framework that combines state-of-the-art machine learning and statistical methodologies. The framework then utilizes a multilayer perceptron (MLP) model to predict pre-transplant survival rates accurately. Using principal component analysis (PCA) for dimensionality reduction provides computational efficiency and emphasizes the most significant features for effectively managing high-dimensional data. Principal component analysis (PCA) reduces the dimensionality of high-dimensional datasets to effectively manage them with computational efficiency and focus more on the most dominant features. Second, we apply association rule mining techniques to identify critical patterns within donor-recipient data. This scalable solution enables dynamic data processing with seamless integration into the clinical workflow. Efficient algorithms are leveraged within the framework to analyze and rank attributes and pinpoint predictors of transplant success. These actionable insights are designed to assist clinicians in making sound decisions, even with the complexity of high-dimensional datasets and improved prediction accuracy. Preliminary results demonstrate that the approach significantly improves donor-recipient matching and survival predictions over existing models. This framework represents a paradigm shift to personalized healthcare and offers a comprehensive mechanism for optimizing liver transplantation outcomes. The algorithms will be validated on larger datasets, and real-time predictive capabilities will be integrated to enhance clinical applications further.