MLP Networks project
This is the project report for the lecture CE434 Neural Networks at Nanyang Technological University, Singapore. Using Java, I explored the use of Multilevel Perceptron (MLP) Networks to approximate a complex nonlinear function and to classify recorded vowel data.
Abstract
Several experiments using MLP networks to approximate a given function and to classify recorded vowel formant frequency data were conducted. This project report presents the background of the experiments, the construction of the networks, the modifications made to enhance performance as well as the results achieved.
ce434-report.pdf (436 KB, 51 pages)
title page - about me - music - computers - people - data - feedback - links - singapore
Copyright © 1996-2003 Christian Renz