. 2020 Jan 7;3:3. doi: 10.1038/s42003-019-0729-3

Table 4.

The Cohen’s kappa agreement scores for automatic analyses, including Deep-channel, SKM, and MDL with golden dataset which is built by five different human experts using their existing software tools.

	Cohen’s kappa score	0.95% CI
MDL	0.766^a	0.7646–0.7674
SKM	0.6497^b	0.6481–0.6513
Deep-channel	0.9279	0.9288–0.9304

^aMDL does not output an idealisation and so we binarized the MDL output using a 50% threshold crossing algorithm to allow comparison

^bSKM implemented in QuB has no true automatic function, one need to set the starting baseline, channel amplitude and number of channels before starting. We did not perform additional baseline correction or input baseline nodes for this comparison