BOOSTING PERFORMANCE IN NEURAL NETWORKS

Harris Drucker; Robert E. Schapire; Patrice Simard

doi:10.1142/s0218001493000352

Abstract

A boosting algorithm, based on the probably approximately correct (PAC) learning model is used to construct an ensemble of neural networks that significantly improves performance (compared to a single network) in optical character recognition (OCR) problems. The effect of boosting is reported on four handwritten image databases consisting of 12000 digits from segmented ZIP Codes from the United States Postal Service and the following from the National Institute of Standards and Technology: 220000 digits, 45000 upper case letters, and 45000 lower case letters. We use two performance measures: the raw error rate (no rejects) and the reject rate required to achieve a 1% error rate on the patterns not rejected. Boosting improved performance significantly, and, in some cases, dramatically.

Keywords

Boosting (machine learning)Artificial neural networkWord error rateComputer scienceArtificial intelligenceGradient boostingOptical character recognitionPostal serviceMachine learningPattern recognition (psychology)Character recognitionImage (mathematics)

Affiliated Institutions

AT&T (United States) US

Related Publications

Improving Performance in Neural Networks Using a Boosting Algorithm

Harris Drucker , Robert E. Schapire , Patrice Simard

A boosting algorithm converts a learning machine with error rate less than 50% to one with an arbitrarily low error rate. However, the algorithm discussed here depends on having...

1992 Neural Information Processing Systems 158 citations

Boosting and Other Ensemble Methods

Harris Drucker , Corinna Cortes , L. D. Jackel +2 more

We compare the performance of three types of neural network-based ensemble techniques to that of a single neural network. The ensemble algorithms are two versions of boosting an...

1994 Neural Computation 352 citations

Handwritten Digit Recognition with a Back-Propagation Network

Yann LeCun , Bernhard E. Boser , John S. Denker +4 more

We present an application of back-propagation networks to handwritten digit recognition. Minimal preprocessing of the data was required, but architecture of the network was high...

1989 neural information processing systems 3629 citations

Comparing support vector machines with Gaussian kernels to radial basis function classifiers

Bernhard Schölkopf , Kah-Kay Sung , Chris Burges +4 more

The Support Vector (SV) machine is a novel type of learning machine, based on statistical learning theory, which contains polynomial classifiers, neural networks, and radial bas...

1997 IEEE Transactions on Signal Processing 1375 citations

Gradient-based learning applied to document recognition

Yann LeCun , Léon Bottou , Yoshua Bengio +1 more

Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate networ...

1998 Proceedings of the IEEE 55110 citations

Publication Info

Year: 1993
Type: article
Volume: 07
Issue: 04
Pages: 705-719
Citations: 184
Access: Closed

External Links

View on DOI.org

Social Impact

Altmetric

BOOSTING PERFORMANCE IN NEURAL NETWORKS

PlumX Metrics

Social media, news, blog, policy document mentions

Citation Metrics

184

OpenAlex

Cite This

APA Style

                            
                                    Harris Drucker, 
                                
                                    Robert E. Schapire, 
                                
                                    Patrice Simard
                                
                            (1993). 
                            BOOSTING PERFORMANCE IN NEURAL NETWORKS. 
                            International Journal of Pattern Recognition and Artificial Intelligence
                            , 07
                            (04)
                            , 705-719.
                            https://doi.org/10.1142/s0218001493000352

Identifiers

DOI: 10.1142/s0218001493000352