Dimensionality reduction using modular perceptron networks

In this paper, we propose a dimension reduction method to overcome the curse of dimensionality on classification problems. The proposed method includes two steps: (1) to create an evaluation function for the reduction criterion, (2) to generate an algorithm for the dimension reduction. We have applied the proposed method on a modular perceptron network (MPN) to learn the real-world datasets. It shows that from the experimental results the dimension of the input data can be decreased largely (less 75% ~ 88%),the data presentations are reduced (less 67% ~ 91%) and a small size MPNs can be procured with learning and testing performance maintained as the good level as before