Comparative study of classification algorithms for immunosignaturing data

BMC Bioinformatics

Table 9 Worst case time performance (in ms) of classification algorithms

Data set	Diabetes	Alzheimer’s	Antibodies	Avg. (in ms)	Rank
Random Tree	1809	491	1478	1260	1
KNN	3016	607	910	1511	2
Hyper Pipes	2486	602	2180	1756	3
Naïve Bayes	4780	1158	2480	2806	4
VFI	7440	1357	3000	3932	5
J48	16581	1385	11731	9899	6
K star	25974	2348	6341	11555	7
SVM	10496	2722	29008	14076	8
R. Forest	50087	8032	21452	26524	9
M5P	50290	8563	23452	27435	10
Bayes Net	55672	9031	25000	29901	11
K-means	85955	12405	29658	42672	12
SLR	632840	48215	605365	428806	13
LDA	658668	869523	632983	720391	14
Logistic R.	1589092	1146783	1315256	1350377	15
ASC	5444533	2465021	4565896	4158483	16
MLP	dnf	dnf	dnf	NA	17

Table showing time performance in milliseconds over >1000 peptides for three datasets. Random Tree, KNN, Hyper Pipes and VFI were among the fastest. MLP were among the slowest with dnf: “Did not finish”. Time measurements less than 10 seconds are marked in bold.

ISSN: 1471-2105