Skip to main content
Fig. 4 | Journal of Translational Medicine

Fig. 4

From: Classification of tumor types using XGBoost machine learning model: a vector space transformation of genomic alterations

Fig. 4

A Bar chart showing the complete overview and detail of the features with the greatest importance based on features ranking of impact on the predicted output, automatically produced by XGBoost during the training of the top 16 tumor types. Abbreviations: Single Nucleotide Polymorphisms (SNPs), Deletions (DELs), Insertions (INSs), CNV Deletions (DLTs), Shallow-Deletions (SHDs), Gains (GANs) and Amplifications (AMPs). B Comparison of SPMs patterns, thus distribution of alterations at chromosome arm-level, between colorectal adenocarcinoma and stomach adenocarcinoma reported as an example of similarity between tumors at molecular level (cross-cancer similarity) and a possible cause of misclassification errors

Back to article page