Credit Risk Prediction Using Fuzzy Immune Learning

Table 3

Predictive accuracy of FAIS and different classifiers.

Classifier	Australian	German
AIRS1: Weka version of [66]	83.2	68.71
AIRS2: Weka version of [66]	84.41	69.64
AIRS as conventional AIS [9]	85.2	71.3
CCS-FAIS [55]	80.7	71.1
CLONALG: Weka version of [59]	82.22	66.42
CSCA*	84.78	70.17
DMNBtext	82.58	69.99
DTNB	85.41	71.52
FAIS [30]	85.51	72
Immunos-1*	76.12	61.38
Immunos-99*	76.72	63.67
LibSVM	85.51	70.98
LWL	85.51	70
Kstar	78.88	69.89
PART	83.32	70.11
SAIS [9]	85.2	75.4
SMO with RBFKernel	85.51	70
IFAIS using simple memory	86.52	72.3
IFAIS using 3-layer memory	87.83	74.9

The results have been obtained using Weka machine learning tool.
The order of classifiers is alphabetical. The most accurate is bold and the second most is italic.
Comparison of predictive accuracies illustrated our proposed algorithm is competitive with other classifiers.
*means classifiers which have been manually added to Weka and are available at http://wekaclassalgos.sourceforge.net.