A regularized stochastic configuration network based on weighted mean of vectors for regression
The stochastic configuration network (SCN) randomly configures the input weights and biases of hidden layers under a set of inequality constraints to guarantee its universal approximation property. The SCN has demonstrated great potential for fast and efficient data modeling. However, the prediction...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
PeerJ Inc.
2023-05-01
|
| Series: | PeerJ Computer Science |
| Subjects: | |
| Online Access: | https://peerj.com/articles/cs-1382.pdf |
| _version_ | 1797814009022382080 |
|---|---|
| author | Yang Wang Tao Zhou Guanci Yang Chenglong Zhang Shaobo Li |
| author_facet | Yang Wang Tao Zhou Guanci Yang Chenglong Zhang Shaobo Li |
| author_sort | Yang Wang |
| collection | DOAJ |
| description | The stochastic configuration network (SCN) randomly configures the input weights and biases of hidden layers under a set of inequality constraints to guarantee its universal approximation property. The SCN has demonstrated great potential for fast and efficient data modeling. However, the prediction accuracy and convergence rate of SCN are frequently impacted by the parameter settings of the model. The weighted mean of vectors (INFO) is an innovative swarm intelligence optimization algorithm, with an optimization procedure consisting of three phases: updating rule, vector combining, and a local search. This article aimed at establishing a new regularized SCN based on the weighted mean of vectors (RSCN-INFO) to optimize its parameter selection and network structure. The regularization term that combines the ridge method with the residual error feedback was introduced into the objective function in order to dynamically adjust the training parameters. Meanwhile, INFO was employed to automatically explore an appropriate four-dimensional parameter vector for RSCN. The selected parameters may lead to a compact network architecture with a faster reduction of the network residual error. Simulation results over some benchmark datasets demonstrated that the proposed RSCN-INFO showed superior performance with respect to parameter setting, fast convergence, and network compactness compared with other contrast algorithms. |
| first_indexed | 2024-03-13T08:01:10Z |
| format | Article |
| id | doaj.art-6c2af6ed9d344a3eabe703cd69c97ee0 |
| institution | Directory Open Access Journal |
| issn | 2376-5992 |
| language | English |
| last_indexed | 2024-03-13T08:01:10Z |
| publishDate | 2023-05-01 |
| publisher | PeerJ Inc. |
| record_format | Article |
| series | PeerJ Computer Science |
| spelling | doaj.art-6c2af6ed9d344a3eabe703cd69c97ee02023-06-01T15:05:04ZengPeerJ Inc.PeerJ Computer Science2376-59922023-05-019e138210.7717/peerj-cs.1382A regularized stochastic configuration network based on weighted mean of vectors for regressionYang Wang0Tao Zhou1Guanci Yang2Chenglong Zhang3Shaobo Li4State Key Laboratory of Public Big Data, Guizhou University, Guiyang, Guizhou, ChinaState Key Laboratory of Public Big Data, Guizhou University, Guiyang, Guizhou, ChinaKey Laboratory of Advanced Manufacturing Technology of the Ministry of Education, Guizhou University, Guiyang, Guizhou, ChinaSchool of Computer Science and Technology, China University of Mining and Technology, Xuzhou, Jiangsu, ChinaState Key Laboratory of Public Big Data, Guizhou University, Guiyang, Guizhou, ChinaThe stochastic configuration network (SCN) randomly configures the input weights and biases of hidden layers under a set of inequality constraints to guarantee its universal approximation property. The SCN has demonstrated great potential for fast and efficient data modeling. However, the prediction accuracy and convergence rate of SCN are frequently impacted by the parameter settings of the model. The weighted mean of vectors (INFO) is an innovative swarm intelligence optimization algorithm, with an optimization procedure consisting of three phases: updating rule, vector combining, and a local search. This article aimed at establishing a new regularized SCN based on the weighted mean of vectors (RSCN-INFO) to optimize its parameter selection and network structure. The regularization term that combines the ridge method with the residual error feedback was introduced into the objective function in order to dynamically adjust the training parameters. Meanwhile, INFO was employed to automatically explore an appropriate four-dimensional parameter vector for RSCN. The selected parameters may lead to a compact network architecture with a faster reduction of the network residual error. Simulation results over some benchmark datasets demonstrated that the proposed RSCN-INFO showed superior performance with respect to parameter setting, fast convergence, and network compactness compared with other contrast algorithms.https://peerj.com/articles/cs-1382.pdfStochastic configuration networksSwarm intelligence optimizationWeighted mean of vectorsResidual error feedback |
| spellingShingle | Yang Wang Tao Zhou Guanci Yang Chenglong Zhang Shaobo Li A regularized stochastic configuration network based on weighted mean of vectors for regression PeerJ Computer Science Stochastic configuration networks Swarm intelligence optimization Weighted mean of vectors Residual error feedback |
| title | A regularized stochastic configuration network based on weighted mean of vectors for regression |
| title_full | A regularized stochastic configuration network based on weighted mean of vectors for regression |
| title_fullStr | A regularized stochastic configuration network based on weighted mean of vectors for regression |
| title_full_unstemmed | A regularized stochastic configuration network based on weighted mean of vectors for regression |
| title_short | A regularized stochastic configuration network based on weighted mean of vectors for regression |
| title_sort | regularized stochastic configuration network based on weighted mean of vectors for regression |
| topic | Stochastic configuration networks Swarm intelligence optimization Weighted mean of vectors Residual error feedback |
| url | https://peerj.com/articles/cs-1382.pdf |
| work_keys_str_mv | AT yangwang aregularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT taozhou aregularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT guanciyang aregularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT chenglongzhang aregularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT shaoboli aregularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT yangwang regularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT taozhou regularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT guanciyang regularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT chenglongzhang regularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression AT shaoboli regularizedstochasticconfigurationnetworkbasedonweightedmeanofvectorsforregression |