Kleve, H. (2004) Using re-entrant mapping in neural network ensembles. Master's Thesis / Essay, Computing Science.
|
Text
Infor_Ma_2004_HKleve.CV.pdf - Published Version Download (1MB) | Preview |
Abstract
Combining multi-modal information can ease classification. Neurobiologist Gerald Edelman postulated a theory on how this kind of information is processed in the brain. He concluded that small local groups of neurons, each from within the context of their own domain, build up connections to each other when activated together. As a result, activation in one area results in activation in connected areas; this mechanism of association can be used for recognition. This study investigates how this neurobiological model can be translated into a computational model and whether combining artificial neural networks this way will improve classification. Currently, neural network ensembles are often employed to combine networks. The ensemble techniques will play a key role in the implementation. An extension of these techniques makes it possible to benefit from unlabelled data. The resulting system shows an improved classification accuracy, with reduced labelled data required for learning.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Degree programme: | Computing Science |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 15 Feb 2018 07:30 |
| Last Modified: | 15 Feb 2018 07:30 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/8945 |
Actions (login required)
 |
View Item |