The week's pick
Random Articles
Reseach Article
Extreme Learning Machine based on Capped ℓ1 Regularization and Pinball Loss Function
| International Journal of Computer Applications |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 187 - Number 1 |
| Year of Publication: 2025 |
| Authors: Zihao Zhou, Jinge Li, Jiayi Jie, Zhenyu Zhu, Xibo Ming |
10.5120/ijca2025924725
|
Zihao Zhou, Jinge Li, Jiayi Jie, Zhenyu Zhu, Xibo Ming . Extreme Learning Machine based on Capped ℓ1 Regularization and Pinball Loss Function. International Journal of Computer Applications. 187, 1 ( May 2025), 25-31. DOI=10.5120/ijca2025924725
Abstract
Extreme Learning Machines (ELM) traditionally employ the squared loss function as the training criterion. However, this function is highly sensitive to outliers, which can amplify their impact on training outcomes of the model, causing the model to deviate from the true data distribution and reducing robustness of the model. Additionally, traditional ELM may encounter overfitting issues when dealing with high-dimensional dataset. To tackle these issues, this study introduces an innovative ELM frameworkthat integrates capped ℓ1 regularization with pinball loss function, termed as Cℓ1-PELM. The capped ℓ1 regularization helps prevent overfitting, and the pinball loss function, due to its linear relationship with the error, effectively mitigates the adverse effects of outliers on model training. This paper employs an iterative reweighting algorithm to optimize the objective function, ensuring rapid convergence of the model during the training process. Experimental results on 18 real-world datasets demonstrate that Cℓ1-PELM exhibits superior robustness, generalization performance, and stability in comparison to other advanced algorithms, particularly in environments with outliers.
References
- Huang G B, Zhu Q Y, Siew C K. 2004Extreme learning machine: a new learning scheme of feedforward neural networks[C]// IEEE international joint conference on neural networks.
- Wang J, Lu S, Wang S H, et al.2022 A review on extreme learning machine[J]. Multimedia Tools and Applications.
- Lal P V, Srilakshmi U, Venkateswarlu D. 2021Face rec-ognition system using extreme learning machines[C]// 5th International Conference on Information Systems and Computer Networks (ISCON).
- Loizidis S, Kyprianou A, Georghiou G E. 2024Electricity market price forecasting using ELM and Bootstrap analysis: A case study of the German and Finnish Day-Ahead markets[J]. Applied Energy.
- Quan R, Liang W, Wang J, et al. 2024An enhanced fault diagnosis method for fuel cell system using a kernel ex-treme learning machine optimized with improved sparrow search algorithm[J]. International Journal of Hydrogen Energy.
- Wang K, Ding X. 2019Pinball loss based extreme learning machines[C]//IOP Conference Series: Materials Science and Engineering.
- Yang J, Cao J, Xue A. 2020Robust maximum mixture correntropy criterion-based semi-supervised ELM with variable center[J]. IEEE Transactions on Circuits and Systems II: Express Briefs.
- Bala R, Dagar A, et al. 2021A novel online sequential extreme learning machine with L2, 1-norm regularization for prediction problems[J]. Applied Intelligence.
- Dai Y, Zhang Y, Wu Q.2024 A pruning extreme learning machine with L 2, 1/2 regularization for multi-dimensional output problems[J]. International Journal of Machine Learning and Cybernetics.
- Nie F, Huo Z, Huang H. 2017 Joint capped norms mini-mization for robust matrix recovery[C]//The 26th Inter-national Joint Conference on Artificial Intelligence.
- Ma J, Yang L, Sun Q. 2020Capped L1-norm distance metric-based fast robust twin bounded support vector machine[J]. Neurocomputing.
- Balasundaram S, Gupta D. 20141-Norm extreme learning machine for regression and multiclass classification using Newton method[J]. Neurocomputing.
- Wang X, Wang K, She Y, et al. 2023Zero-norm ELM with non-convex quadratic loss function for sparse and robust regression[J]. Neural Processing Letters.
- Anand P, Rastogi R, Bharti A. 2018A Pinball loss based Extreme Learning Machine for pattern classifica-tion[C]IEEE Symposium Series on Computational Intel-ligence.
- Martínez-Martínez J M, 2011Escandell-Montero P, Soria-Olivas E, et al. Regularized extreme learning machine for regression problems[J]. Neurocomputing.
- Deng W, Zheng Q, Chen L. 2009 Regularized extreme learning machine[C] IEEE symposium on computational intelligence and data mining.
- Chen K, Lv Q, Lu Y, 2017et al. Robust regularized ex-treme learning machine for regression using iteratively reweighted least squares[J]. Neurocomputing.
Index Terms
Keywords