next up previous
Next: Conclusion Up: A Modular Neural Previous: Experiments


The network is less useful for problems with very small input dimensions. The network has the ability to solve problems which are not linear separable. The proposed architecture has certain theoretical limitations; statistically neutral problems (like the XOR-problem) can not be learned. Monolithic MLPs are able to learn such problems but the generalization performance is very poor [7].

Albrecht Schmidt and Zuhair Bandar
Wed Apr 16 13:57:11 MET DST 1997