Since decision trees and neural networks are alternative paradigms for classification, much research has addressed the issue of developing hybrid learning algorithms that combine
them together. A review on this topic can be found in [9]. Different to previous hybrid learning algorithms, neither does NeC4.5 use decision tree to help determine the topology of neural networks, nor does it use neural network ensemble to refine the splits or even embed as splits into the decision trees. Therefore NeC4.5 exhibits a new way to hybrid learning.