A feed-forward neural network with one hidden layer and full connection is analyzed from several points of view and trained with different algorithms. A comparative performance analysis shows marginal differences supporting the conclusion that the network architecture has a stronger influence upon the performance than the training algorithm.
Citation:
Claudio Moraga, Adrian Gasch, "Comparative Analysis of Different Learning Algorithms on a Common Neural Network Structure," sccc, pp.19, 19th International Conference of the Chilean Computer Science Society, 1999