As part of my doctoral research, I am engaged in the analysis of cancer registry data using machine learning (ML). The objective is to enhance the predictive power of medical events (survival, recurrence, progression) through the application of machine learning (ML). The ongoing work includes the conversion of categorical variables into numerical representations. Medical classifications, such as the TNM system, grading, or morphology, are essential for determining the stage of a cancer. However, these classifications are frequently unordered or ordered categorical features, whereas machine learning algorithms only process numerical values. However, the process of converting categories into numerical representations often results in the misrepresentation of those information. Accordingly, my doctoral studies concentrate on the development and implementation of ML-approaches for the goal of a more precise and accurate representation of cancer patient data within the latent feature space through numeric representations. This improved representation of oncological patient data is relevant. Firstly, clustering methods are more effective at identifying similar cancers, particularly in rare diseases, given that the quality of the clustering results depends largely on the selection of similarity measures and distance metrics. Secondly, an enhanced representation of cancer registry data in latent space enables more precise prediction of medical events in cancer through the application of ML, as it allows ML algorithms to more accurately identify structures and patterns within the data. Furthermore, improved visualization of cancer registry data facilitates the explanation of ML models (XAI), thereby enhancing their acceptance and trust. Through interdisciplinary collaboration and the exchange of medical experts and data analysts, innovative solutions for the analysis and interpretation of cancer registry data using ML can be developed.