Fig. 4

a: Pictorial representation of Cliff’s delta values when comparing features of shape, histogram and textures across mutation status and grades. Exact Cliff delta values are given in Table S1. Positive values are shown in darkening red while negative ones are in darkening blue. b: Outcomes of the confusion (or error) matrix aiding visualisation of the performance of our machine-learning algorithm. The first row shows the confusion matrices for the classification of gliomas by IDH mutation status. The second row shows the confusion matrices for the classification across the three WHO grades II, III and IV. In both classification scenarios we show three cases: using all data and separately using only data obtained from scanners with 1.5 T or 3 T magnetic field. Within each matrix, the matrix row represent the instances in the actual ground truth class while each column represents the instances in the predicted class. Darkening red correspond to higher percentages of the overall population to be classified. The exact values of the confusion matrices are given in supplementary table S4 of the Appendix.