Table 2.

Methods of comparison between BP-NN, RBF-NN, and PSO-KSVM.

Method	Confusion matrix	Success Cases	Sensitivity	Specificity	Classification accuracy
BP-NN	$\begin{array}{cc}\mathrm{374}& \mathrm{11}\end{array}\hspace{0.17em}$  $\begin{array}{cc}51& 14\end{array}$	388	88.0%	56%	86.22%
RBF-NN	$\begin{array}{cc}393& 7\end{array}\hspace{0.17em}$  $\begin{array}{cc}32& 18\end{array}$	411	92.47%	72%	91.33%
PSO-KSVM	$\begin{array}{cc}417& 2\end{array}\hspace{0.17em}\hspace{0.17em}$  $\hspace{0.17em}\begin{array}{cc}8& 23\end{array}$	440	98.12%	92%	97.78%