Paper Title
Neural Network Architecture through Data Mining Techniques
Abstract
The paper presents the technique used to determine the optimal neural network architecture by means of Pattern
Recognition and Data Mining Techniques. Using clustering techniques on the data to be processed in the context of the way to
work unsupervised and to correlate the number of groups obtained with the optimal number of hidden layers for a multilayer
neural network.
Index Terms—Clustering Methods, Data Mining, Hidden Layers, Neural Network Architecture, Pattern Recognition.