Effectiveness of different target coding schemes on networks in financial engineering

Authors:
Kidong Lee;Junghee Park;Sangjae Lee
Affiliations:
Department of Business, University of Incheon, Incheon, South Korea;School of Information Technology, Jaeneung College, Incheon, South Korea;Department of E-business, College of Business, Sejong University, South Korea
Venue:
ISNN'05 Proceedings of the Second international conference on Advances in neural networks - Volume Part II
Year:
2005

Citing 3
Cited 0

Optimized Feature Extraction and the Bayes Decision in Feed-Forward Classifier Networks

IEEE Transactions on Pattern Analysis and Machine Intelligence
Managerial applications of neural networks: the case of bank failure predictions

Management Science
Bankruptcy prediction using neural networks

Decision Support Systems - Special issue on neural networks for decision support

Quantified Score

Hi-index	0.00

Visualization

Abstract

The effectiveness of backpropagation (BP) network relies mainly on a proper architectural design as well as finding appropriate parameter values by training the network at the same time. In this experiment, we test the effectiveness of three different target coding schemes including Lowe and Webb's method of BP using unmatched business samples in the context of financial engineering. The results of the study show that Lowe and Webb's reciprocally weighted target-coding scheme may not be the best choice among the three target coding schemes in neural network applied. Rather, in neural net experiment, no knowledge about prior proportion on sample shows the promising result.