A bag-of-words approach for Drosophila gene expression pattern annotation

BMC Bioinformatics

Table 5 Annotation performance in terms of AUC, macro F1, micro F1, sensitivity, and specificity for image groups in stage range 11–12.

Measure	# of terms	ML_LS	SVM	PMK_SIFT	PMK_comp
AUC	10	84.24 ± 0.54	83.05 ± 0.54	79.41 ± 0.58	78.68 ± 0.58
	20	84.23 ± 0.37	82.73 ± 0.38	77.62 ± 0.48	76.97 ± 0.67
	30	81.32 ± 0.44	79.18 ± 0.51	73.26 ± 0.42	72.90 ± 0.63
	40	75.68 ± 0.68	77.52 ± 0.63	72.96 ± 0.98	72.12 ± 0.64
	50	77.87 ± 0.81	76.19 ± 0.72	71.55 ± 1.14	70.73 ± 0.83
macro F1	10	61.07 ± 0.99	60.37 ± 0.88	47.22 ± 1.44	55.19 ± 0.62
	20	48.06 ± 0.80	46.03 ± 0.83	26.36 ± 0.85	36.11 ± 0.88
	30	34.92 ± 0.99	35.32 ± 0.75	18.00 ± 0.64	24.85 ± 0.63
	40	26.14 ± 0.65	27.85 ± 0.79	13.31 ± 0.47	19.01 ± 0.63
	50	22.77 ± 1.00	23.46 ± 0.60	10.79 ± 0.41	15.04 ± 0.46
micro F1	10	66.77 ± 0.72	65.67 ± 0.60	60.41 ± 0.93	62.55 ± 0.68
	20	59.83 ± 0.65	54.59 ± 0.87	50.49 ± 0.71	53.61 ± 0.59
	30	54.99 ± 0.79	48.87 ± 0.85	47.01 ± 0.73	49.61 ± 0.64
	40	49.49 ± 0.71	46.40 ± 0.91	45.38 ± 0.61	48.10 ± 0.75
	50	52.15 ± 0.67	47.18 ± 0.84	44.95 ± 0.56	47.41 ± 0.61
Sensitivity	10	66.99 ± 1.17	70.18 ± 1.78	46.92 ± 1.44	64.43 ± 0.96
	20	48.35 ± 1.31	55.43 ± 1.56	24.81 ± 0.69	43.39 ± 1.22
	30	34.80 ± 1.33	48.99 ± 1.87	16.64 ± 0.45	31.00 ± 0.97
	40	26.96 ± 1.20	37.42 ± 1.33	12.35 ± 0.32	23.49 ± 0.96
	50	22.78 ± 1.11	30.36 ± 1.24	10.01 ± 0.30	18.68 ± 0.69
Specificity	10	80.84 ± 1.19	77.58 ± 1.29	84.11 ± 0.86	80.93 ± 0.78
	20	92.00 ± 0.52	87.85 ± 0.78	93.93 ± 0.30	92.76 ± 0.29
	30	94.33 ± 0.40	88.67 ± 0.92	96.38 ± 0.19	95.52 ± 0.21
	40	94.80 ± 0.43	92.10 ± 0.56	97.30 ± 0.14	96.65 ± 0.15
	50	96.29 ± 0.25	93.80 ± 0.45	97.85 ± 0.10	97.28 ± 0.12

ISSN: 1471-2105