Skip to main content

Table 2 Median F1 of the eight scRNAseq data classifier tools and the four conventional classifiers on the seven datasets (N = 1000)

From: Single-cell classification using graph convolutional networks

Methods Zhengsorted Zheng68K BaronHuman Muraro Segerstolpe BaronMouse Xin
sigGCN 0.965 0.776 0.977 1 1 0.969 0.995
FC 0.952 0.681 0.938 1 0.968 0.902 0.997
scID 0.66 0.535 0.22 0.578 0 0 1
scPred 0.568 0.105 0.833 0.932 0.8 0.97 0.784
CasTLe 0.834 0.667 0.956 0.967 0.965 0.848 0.997
SingleR 0.678 0.335 0.946 0.984 1 0.898 1
scmapcluster 0.729 0.357 0.9 0.997 0.965 0.88 0.991
scmapcell 0.305 0.198 0.95 0.993 0.977 0.942 0.708
ACTINN 0.892 0.753 1 0.97 1 1 0.997
RF 0.853 0.646 0.956 0.987 0.993 0.984 0.997
SVM-linear 0.868 0.663 0.362 1 0.059 0.238 1
SVM-rbf 0.9 0.671 0.906 1 0.772 0.695 1
KNN 0.795 0.613 0.928 1 0.961 0.889 1