Skip to main content
. 2017 Jul 13;33(21):3423–3430. doi: 10.1093/bioinformatics/btx448

Table 1.

Summary of results for the FlowCAP-I cell identification challenge

Dataset DeepCyTOF Competition’s winner
GvHD 0.986 (0.979, 0.991) 0.92 (0.88, 0.95)
DLBCL 0.985 (0.976, 0.993) 0.95 (0.93, 0.97)
HSCT 0.991 (0.988, 0.993) 0.98 (0.96, 0.99)
WNV 0.999 (0.998, 0.999) 0.96 (0.94, 0.97)
ND 0.988 (0.987, 0.989) 0.94 (0.92, 0.95)

Note: The numbers in parentheses represent 95% confidence intervals for the F-measure. We use DeepCyTOF without denoising (DAE) and without calibration (MMD-ResNets).