Conclusions

• Proved comparable to existing rule generation algorithms such as C5.0, ID3 and the RULES
• Particularly wrt the size and accuracy of the rule set created.
• Has shown Immune-based methods suitable for rule extraction