Skip to main content

Table 11 Summary of performance and time measures of classification algorithms

From: Comparative study of classification algorithms for immunosignaturing data

 

# Rank 1

# Rank 2

# >90%

Distance

Time

Naïve Bayes

5

1

6

−0.3

2X

MLP

0

0

4

−3.4

7615X

SVM

0

1

3

−3.6

11X

VFI

0

2

4

−5.7

3X

Hyper Pipes

0

0

0

−7.9

1X

R. Forest

1

0

2

−8.8

21X

Bayes Net

0

1

2

−8.8

24X

K-means

0

0

1

−9.9

34X

Logistic R.

0

1

3

−11.8

1072X

SLR

1

1

2

−12.9

340X

KNN

0

0

1

−14.4

1X

K star

0

0

1

−16.5

9X

M5P

0

0

0

−17.0

22X

J48

0

0

0

−20.2

8X

Random Tree

0

0

0

−20.7

1X

ASC

0

0

0

−22.1

3300X

LDA

0

0

0

−24.0

572X

  1. #Rank 1, Rank 2: No. of times algorithm ranked 1st and 2nd on 7 datasets, # > 90%: No. of times algorithm scored overall average score >90% on 7 datasets, Distance: magnitude an algorithm trails behind on average from the Rank 1 for the datasets (5% or less distance are marked in bold). Time: performance slower with respective to fastest algorithm. Time performances slower by 5 folds to fastest algorithm are marked in bold.