Gap-Based Estimation: Choosing the Smoothing Parameters for Probabilistic and General Regression Neural Networks

Authors:
Mingyu Zhong;Dave Coggeshall;Ehsan Ghaneie;Thomas Pope;Mark Rivera;Michael Georgiopoulos;Georgios C. Anagnostopoulos;Mansooreh Mollaghasemi;Samuel Richie
Affiliations:
myzhong@ucf.edu;david.coggeshall@gmail.com;Ehsan.Ghaneie@gmail.com;ThomasPope@gmail.com;mark.rivera@gmail.com;School of Electrical Engineering and Computer Science, University of Central Florida, Orlando, FL 32816, U.S.A. michaelg@mail.ucf.edu;Department of Electrical and Computer Engineering, Florida Institute of Technology, Melbourne, FL 32901, U.S.A. georgio@fit.edu;Department of Industrial Engineering and Management Systems, University of Central Florida, Orlando, FL 32816, U.S.A. mollagha@mail.ucf.edu;School of Electrical Engineering and Computer Science, University of Central Florida, Orlando, FL 32816, U.S.A. richie@mail.ucf.edu
Venue:
Neural Computation
Year:
2007

Citing 5
Cited 1

Integrating neural networks with influence diagrams for multiple sensor diagnostic systems

Integrating neural networks with influence diagrams for multiple sensor diagnostic systems
Gaussian ARTMAP: a neural network for fast incremental learning of noisy multidimensional maps

Neural Networks
A constructive, incremental-learning network for mixture modeling and classification

Neural Computation
Optimization of the kernel functions in a probabilistic neural network analyzing the local pattern distribution

Neural Computation
LIBSVM: A library for support vector machines

ACM Transactions on Intelligent Systems and Technology (TIST)

Parallelism, localization and chain gradient tuning combinations for fast scalable probabilistic neural networks in data mining applications

SETN'12 Proceedings of the 7th Hellenic conference on Artificial Intelligence: theories and applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

Probabilistic neural networks (PNN) and general regression neural networks (GRNN) represent knowledge by simple but interpretable models that approximate the optimal classifier or predictor in the sense of expected value of the accuracy. These models require the specification of an important smoothing parameter, which is usually chosen by cross-validation or clustering. In this article, we demonstrate the problems with the cross-validation and clustering approaches to specify the smoothing parameter, discuss the relationship between this parameter and some of the data statistics, and attempt to develop a fast approach to determine the optimal value of this parameter. Finally, through experimentation, we show that our approach, referred to as a gap-based estimation approach, is superior in speed to the compared approaches, including support vector machine, and yields good and stable accuracy.