Enhanced extreme learning machine via competitive learning SSA (CL-SSA) for load capacity factor prediction
Extreme Learning Machine (ELM) is known for its fast training speed and simplicity of implementation; however, it suffers from certain limitations, including sensitivity to random initialization and inadequate weight optimization, which can result in suboptimal accuracy and precision. This study int...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-01-01
|
| Series: | Heliyon |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405844025002725 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Extreme Learning Machine (ELM) is known for its fast training speed and simplicity of implementation; however, it suffers from certain limitations, including sensitivity to random initialization and inadequate weight optimization, which can result in suboptimal accuracy and precision. This study introduces an enhanced Competitive Learning Salp Swarm Algorithm (CLSSA), which integrates the Salp Swarm Algorithm (SSA) with Competitive Swarm Optimization (CSO) to improve the exploitation capabilities of the traditional CSO. The goal is to address the limitations of traditional ELM by optimizing the weights and biases of the network more effectively, thereby improving the precision and convergence speed of ELM. The research first evaluates the efficiency of the improvement made to the CLSSA optimizer in comparison with various optimization methods, using CEC 2015 benchmark functions to demonstrate the effectiveness of the proposed improvements. The results show that CLSSA outperforms other optimizers in 86 % of the CEC 2015 functions, underscoring its superior optimization capabilities. Furthermore, the study assesses the effectiveness of the CLSSA-enhanced ELM (ELM-CLSSA) in predicting the load capacity factor. The findings reveal that the hybrid ELM-CLSSA framework significantly outperforms both alternative approaches and the traditional ELM framework in terms of training and prediction accuracy, achieving an impressive accuracy rate of 97%. The algorithm's rapid convergence, high precision, and ability to avoid local optima make it a promising solution for complex problems, such as load capacity factor prediction, which is critical for environmentally sustainable initiatives. In addition, the feature analysis conducted by ELM-CLSSA provides valuable insights into the key variables influencing load capacity factor prediction, highlighting the importance of factors such as coal energy, economic growth, technological innovation, and biomass. This study advocates for the use of the ELM-CLSSA framework to improve the precision and reliability of load capacity factor prediction, offering a valuable tool for scientists and policymakers in their efforts to promote ecological conservation and combat climate change. |
|---|---|
| ISSN: | 2405-8440 |