Fig. 2From: Classification of tumor types using XGBoost machine learning model: a vector space transformation of genomic alterationsConfusion matrices showing the performance, in terms of accuracy (ACC), balanced accuracy (BACC) and AUC score, of XGBoost model trained with the 16 (A) and 10 (B) most represented cancer types of the dataset, corresponding to the 70% and 50% of the entire dataset, respectively. The number of calls for somatic point mutations (SPMs) and copy number variations (CNVs) at chromosome arm-level were considered as predictive features of cancer types. This figure shows the two models that achieved the best performanceBack to article page