Skip to main content

Table 2 Comparison of classification performance of the parallel and sequential GA with other classification algorithms |x| = 40.

From: A parallel genetic algorithm for single class pattern classification and its application for gene expression profiling in Streptomyces coelicolor

Pattern

Sequent GA

Parallel GA (2 islands)

Parallel GA (4 islands)

Binary SVM

Single SVM

LogitBoost

LR

LDA

LS

 

Se

Sp

Se

Sp

Se

Sp

Se

Sp

Se

Sp

Se

Sp

Se

Sp

Se

Sp

Se

Sp

01

1

0.9786

1

0.9862

1

0.9922

1

0.9884

0.925

0.959

1

0.4906

1

0.6602

1

0.2076

0.9625

0.6234

02

1

0.9836

1

0.9858

1

0.9866

1

0.9556

0.8375

0.9918

0.9875

0.6846

1

0.326

1

0.2152

0.975

0.8174

03

1

0.9928

1

0.995

1

0.9972

1

0.9844

0.8875

0.8452

-

-

1

0.7042

1

0.3732

0.3375

0.3866

  1. Sequential GA, parallel GA (2 and 4 islands mode) and support vector machines (SVM binary and single), logitBoost, linear discriminant analysis (LDA), logistic regression (LR), and linear least squares regression (LS) for three sets of template vectors of different dimensions were tested. Se and Sp are defined in Eq. 7 and 8.