| dc.contributor.author |
Glezakos, TJ |
en |
| dc.contributor.author |
Tsiligiridis, TA |
en |
| dc.contributor.author |
Yialouris, CP |
en |
| dc.date.accessioned |
2014-06-06T06:53:06Z |
|
| dc.date.available |
2014-06-06T06:53:06Z |
|
| dc.date.issued |
2014 |
en |
| dc.identifier.issn |
09410643 |
en |
| dc.identifier.uri |
http://dx.doi.org/10.1007/s00521-012-1212-y |
en |
| dc.identifier.uri |
http://62.217.125.90/xmlui/handle/123456789/6370 |
|
| dc.subject |
Artificial neural networks |
en |
| dc.subject |
Evolutionary computing |
en |
| dc.subject |
Machine learning |
en |
| dc.subject |
Plant virus identification |
en |
| dc.subject |
Support vector machines |
en |
| dc.subject |
Torrential risk management |
en |
| dc.subject.other |
Evolutionary computing |
en |
| dc.subject.other |
Plant virus |
en |
| dc.subject.other |
Secondary data sets |
en |
| dc.subject.other |
Time series data analysis |
en |
| dc.subject.other |
Time series informations |
en |
| dc.subject.other |
Time series modeling |
en |
| dc.subject.other |
Time series models |
en |
| dc.subject.other |
Time-series segmentation |
en |
| dc.subject.other |
Feature extraction |
en |
| dc.subject.other |
Learning systems |
en |
| dc.subject.other |
Neural networks |
en |
| dc.subject.other |
Risk assessment |
en |
| dc.subject.other |
Risk management |
en |
| dc.subject.other |
Support vector machines |
en |
| dc.subject.other |
Viruses |
en |
| dc.subject.other |
Time series |
en |
| dc.title |
Piecewise evolutionary segmentation for feature extraction in time series models |
en |
| heal.type |
journalArticle |
en |
| heal.identifier.primary |
10.1007/s00521-012-1212-y |
en |
| heal.publicationDate |
2014 |
en |
| heal.abstract |
The design, development and implementation of an innovative system utilized in feature extraction from time series data models is described in this manuscript. Achieving to design piecewise segmentation patterns on the time series in an evolutionary fashion and use them in order to produce fitter secondary data sets, the developed system adapts itself to the nature of the problem each time and finally elects an optimally parameterized classifier (artificial neural network or support vector machine), along with the fittest time series segmentation pattern. The application of the system onto two different problems involving time series data analysis and requiring predictive and classification capabilities (torrential risk assessment and plant virus identification, respectively), reveals that the proposed methodology was crucial in finding the optimum solution for both problems. Piecewise evolutionary segmentation time series model analysis, utilized by the accompanying software tool, succeeded in controlling the dimensionality and noise inherent in the initial raw time series information. The process eventually proposes a segmentation pattern for each problem, enhancing the potential of the corresponding classifier. © 2012 Springer-Verlag London. |
en |
| heal.journalName |
Neural Computing and Applications |
en |
| dc.identifier.issue |
2 |
en |
| dc.identifier.volume |
24 |
en |
| dc.identifier.doi |
10.1007/s00521-012-1212-y |
en |
| dc.identifier.spage |
243 |
en |
| dc.identifier.epage |
257 |
en |