A method to improve protein subcellular localization prediction by integrating various biological data sources

BMC Bioinformatics

Table 2 Performance comparison between fuzzy k-NN and k-NN models in three measures

	ISORT (1-N)	ISORT (2-NN)	ISORT (3-NN)	Fuzzy K-NN (k = 25, m = 1.05)
Measure I (k = 1) (%)	50.68	55.41	56.91	62.25
Measure I (k = 2) (%)	59.67	68.85	70.40	79.77
Measure I (k = 3) (%)	60.23	72.93	76.96	86.14
Measure II (%)	47.83	55.73	58.63	63.52
Mitochondrion	43.81	28.43	38.13	35.12
Vacuole	30.26	26.32	26.32	31.58
Spindle pole	27.78	16.67	22.22	38.89
Cell periphery	26.98	31.75	30.16	34.92
Punctate composite	6.56	4.92	3.28	19.67
Vacuolar membrane	8.11	10.81	0	8.11
ER	41.61	44.97	41.61	53.02
Nuclear periphery	50.00	35.00	45.00	50.00
Endosome	40.74	40.74	40.74	40.74
Bud neck	36.11	30.56	33.33	36.11
Microtubule	45.46	45.46	45.46	45.46
Golgi	28.57	28.57	23.81	23.81
Late Golgi	21.74	13.04	17.39	21.74
Peroxisome	33.33	33.33	33.33	33.33
Actin	52.94	23.53	23.53	52.94
Nucleolus	13.92	15.19	20.25	32.91
Cytoplasm	49.08	64.72	66.18	79.88
ER to Golgi	100.00	100.00	100.00	100.00
Early Golgi	20.00	30.00	33.33	26.67
Lipid particle	18.18	9.09	27.27	9.09
Nucleus	63.47	78.03	83.25	77.91
Bud	76.92	53.85	23.08	7.69
Measure III (%)	37.98	34.77	35.35	39.07

ISSN: 1471-2105