The Radiomics area concerns using medical image features to predict biological phenomenas. For example, we try to predict tumor phenotype using features extracted from MR and CT scans. A complete Radiomics workflow consists of multiple individual parts. These span several technical areas, such as image segmentation, registration, feature extractions and classification. For each step, a wide variety of algorithms exists. Different combinations of these might be optimal for different datasets and prediction problems. As these areas are widely covered and ever growing, testing of all different options is nearly impossible due to the high computation time. The aim of this project is therefore to find a method that will automatically select the best algorithm combination for each problem. In machine learning, the field of automatic learning algorithms is known as meta-learning. We want to extend this to include the image processing areas mentioned above and find an efficient and intuitive solution.