Peak intensity prediction in MALDI-TOF mass spectrometry: A machine learning study to support quantitative proteomics

Table 2 Overview of Pearson's correlation coefficients using mic normalization

validation	dataset	feature set	ν-SVR	LLM	LM
10-fold CV	A	aa	0.66	0.52	0.60
		sss	0.66	0.67	0.51
		mono	0.68	0.64	0.52
		seq	0.57	0.34	0.34
	B	aa	0.53	0.46	0.49
		sss	0.55	0.53	0.49
		mono	0.47	0.53	0.48
		seq	0.44	0.27	0.41
across datasets	A	aa	0.65	0.52	0.14
		sss	0.63	0.59	0.47
		mono	0.66	0.57	0.45
		seq	0.46	0.21	0.40
	B	aa	0.45	0.24	0.01
		sss	0.50	0.44	0.39
		mono	0.45	0.39	0.32
		seq	0.32	0.05	0.28
	A	aa	0.58	0.47	0.00
		sss	0.58	0.55	0.41
		mono	0.61	0.52	0.39
across datasets		seq	0.37	0.21	0.22
without duplicates	B	aa	0.44	0.42	0.00
		sss	0.53	0.46	0.40
		mono	0.46	0.44	0.32
		seq	0.32	0.00	0.03

Values "0.00" indicate that the correlation coefficient was in the range (-0.005,0.005). The best value in each section is printed in bold face.

ISSN: 1471-2105