Table 1

Overall performance measure of classification algorithms on datasets
Algorithms T1D Az Ab Asthma A & B A & C B & D Avg. Rank
Naïve Bayes 92.0 93.4 91.5 77.7 90.8 93.5 93.6 90.4 1
MLP 90.1 92.7 90.2 71.1 84.7 92.7 89.3 87.3 2
SVM 91.6 88.0 90.7 71.3 86.1 88.4 93.1 87.0 3
VFI 90.5 92.2 75.5 62.6 87.7 93.4 92.7 84.9 4
Hyper Pipes 89.8 89.7 81.3 62.3 82.0 86.6 87.8 82.8 5
R. Forest 91.5 82.4 93.3 62.8 80.6 81.4 81.1 81.9 6
Bayes Net 90.3 87.7 92.5 53.9 80.2 83.2 85.1 81.8 7
K-means 88.3 91.8 80.7 59.6 77.8 83.3 83.6 80.7 8
Logistic R. 90.6 93.3 60.4 50.7 81.5 84.8 90.7 78.9 9
SLR 92.2 71.8 90.1 72.2 65.0 68.5 84.7 77.8 10
KNN 91.4 81.5 52.5 55.8 87.5 75.7 89.0 76.2 11
K star 81.9 90.7 89.4 53.5 64.3 68.8 70.7 74.2 12
M5P 85.1 58.7 83.2 60.0 75.2 73.4 79.6 73.6 13
J48 80.3 69.7 78.4 48.7 70.6 68.4 76.7 70.4 14
Random Tree 83.8 71.7 76.2 52.9 69.3 60.8 75.0 70.0 15
ASC 76.8 70.0 77.9 43.1 72.0 63.1 76.7 68.5 16
LDA 69.7 52.0 89.1 70.8 62.8 69.7 52.6 66.7 17

T1D: Type 1 diabetes datasets, Az: Alzehemer’s dataset, Ab: Antibodies dataset. Table showing algorithms overall performance in each datasets based on average score. Score >90% are marked in bold. Naïve Bayes scored the overall highest average score of 90.4%.

Kukreja et al.

Kukreja et al. BMC Bioinformatics 2012 13:139   doi:10.1186/1471-2105-13-139

Open Data