Network ensemble and constructive algorithms for model selection of extreme learning machine
The extreme learning machine (ELM) introduced by Huang et al. is a learning algorithm designed based on the generalized SLFNs with a wide variety of hidden nodes. It randomly generates hidden node parameters and then determines the output weights analytically. ELM is very simple and it tends to obta...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/44760 |
| _version_ | 1811690460629958656 |
|---|---|
| author | Lan, Yuan |
| author2 | Huang Guangbin |
| author_facet | Huang Guangbin Lan, Yuan |
| author_sort | Lan, Yuan |
| collection | NTU |
| description | The extreme learning machine (ELM) introduced by Huang et al. is a learning algorithm designed based on the generalized SLFNs with a wide variety of hidden nodes. It randomly generates hidden node parameters and then determines the output weights analytically. ELM is very simple and it tends to obtain the smallest training error and the smallest norm of weights, which can lead to good generalization performance of networks. However, the good performance of ELM is valid only when the network architecture is chosen correctly. This thesis investigated the problems of network architecture design and model selection of ELM. Essentially, in the thesis, we proposed the use of network ensemble to improve the generalization performance of online ELM network and then we focused on the novel constructive approaches to alter the network structure during the learning process in order to find the appropriate architecture. A parsimonious structure can be found by the constructive method with a backward refinement phase. |
| first_indexed | 2024-10-01T06:04:21Z |
| format | Thesis |
| id | ntu-10356/44760 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T06:04:21Z |
| publishDate | 2011 |
| record_format | dspace |
| spelling | ntu-10356/447602023-07-04T16:54:35Z Network ensemble and constructive algorithms for model selection of extreme learning machine Lan, Yuan Huang Guangbin Soh Yeng Chai School of Electrical and Electronic Engineering DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems The extreme learning machine (ELM) introduced by Huang et al. is a learning algorithm designed based on the generalized SLFNs with a wide variety of hidden nodes. It randomly generates hidden node parameters and then determines the output weights analytically. ELM is very simple and it tends to obtain the smallest training error and the smallest norm of weights, which can lead to good generalization performance of networks. However, the good performance of ELM is valid only when the network architecture is chosen correctly. This thesis investigated the problems of network architecture design and model selection of ELM. Essentially, in the thesis, we proposed the use of network ensemble to improve the generalization performance of online ELM network and then we focused on the novel constructive approaches to alter the network structure during the learning process in order to find the appropriate architecture. A parsimonious structure can be found by the constructive method with a backward refinement phase. DOCTOR OF PHILOSOPHY (EEE) 2011-06-03T07:37:18Z 2011-06-03T07:37:18Z 2011 2011 Thesis Lan, Y. (2011). Network ensemble and constructive algorithms for model selection of extreme learning machine. Doctoral thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/44760 10.32657/10356/44760 en 182 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems Lan, Yuan Network ensemble and constructive algorithms for model selection of extreme learning machine |
| title | Network ensemble and constructive algorithms for model selection of extreme learning machine |
| title_full | Network ensemble and constructive algorithms for model selection of extreme learning machine |
| title_fullStr | Network ensemble and constructive algorithms for model selection of extreme learning machine |
| title_full_unstemmed | Network ensemble and constructive algorithms for model selection of extreme learning machine |
| title_short | Network ensemble and constructive algorithms for model selection of extreme learning machine |
| title_sort | network ensemble and constructive algorithms for model selection of extreme learning machine |
| topic | DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems |
| url | https://hdl.handle.net/10356/44760 |
| work_keys_str_mv | AT lanyuan networkensembleandconstructivealgorithmsformodelselectionofextremelearningmachine |