Skip to main content

Table 2 Parameter estimation: The performance of weighted SVMs with different kernel functions, λ = 1.5 and β = 0.025.

From: Contextual weighting for Support Vector Machines in literature mining: an application to gene versus protein name disambiguation

Kernel

AUC

Parameters

Linear

85.31%

C = 1

Gaussian

86.15%

C = 1, γ = 1

Polynomial (d = 2)

86.09%

C = 2-3, γ = 2

Polynomial (d = 3)

86.13%

C = 2-3, γ = 1