File(s) under permanent embargo
Improving prediction interval quality : a genetic algorithm-based method applied to neural networks
journal contribution
posted on 2009-12-15, 00:00 authored by Abbas KhosraviAbbas Khosravi, Saeid Nahavandi, Douglas CreightonDouglas CreightonThe delta technique has been proposed in literature for constructing
prediction intervals for targets estimated by neural networks. Quality of constructed prediction intervals using this technique highly depends on neural network characteristics. Unfortunately, literature is void of information about how these dependences can be managed in order to optimize prediction intervals. This study attempts to optimize length and coverage probability of prediction intervals through modifying structure and parameters of the underlying neural networks. In an evolutionary optimization, genetic algorithm is applied for finding the optimal values of network size and training hyper-parameters. The applicability and efficiency of the proposed optimization technique is examined and demonstrated using a real case study. It is shown that application of the proposed optimization technique significantly improves quality of constructed prediction intervals in term of length and coverage probability.
prediction intervals for targets estimated by neural networks. Quality of constructed prediction intervals using this technique highly depends on neural network characteristics. Unfortunately, literature is void of information about how these dependences can be managed in order to optimize prediction intervals. This study attempts to optimize length and coverage probability of prediction intervals through modifying structure and parameters of the underlying neural networks. In an evolutionary optimization, genetic algorithm is applied for finding the optimal values of network size and training hyper-parameters. The applicability and efficiency of the proposed optimization technique is examined and demonstrated using a real case study. It is shown that application of the proposed optimization technique significantly improves quality of constructed prediction intervals in term of length and coverage probability.
History
Journal
Lecture notes in computer scienceVolume
5864Pagination
141 - 149Publisher
SpringerLocation
Heidelberg, GermanyPublisher DOI
ISSN
0302-9743eISSN
1611-3349Language
engPublication classification
C1 Refereed article in a scholarly journalCopyright notice
2009, Springer-Verlag Berlin HeidelbergUsage metrics
Categories
No categories selectedKeywords
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC