.png){kind=icon}
.png){kind=icon}
Return
Efficient Large-Scale Classification with Linex Least Square Twin Bounded Support Vector Machine
.png){kind=icon}
PRE
AI
DOI：10.37193/CJM.2026.02.06.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper addresses the challenges of large-scale classification problems in machine learning, where traditional algorithms often struggle with computational and memory constraints, resulting in prolonged training times. To overcome these limitations, we propose LSTBSVM-linex, a model that combines an asymmetric Linex loss function with Least Square Twin Bounded Support Vector Machine (LSTBSVM), optimized for large-scale datasets using the Adam algorithm. Our proposed model demonstrates superior performance in terms of both classification robustness and computational efficiency, as supported by convergence analysis and numerical results. Additionally, statistical tests are employed to validate the competitive performance of LSTBSVM-linex. The effectiveness of the model is further illustrated through its successful application to Knee Osteoarthritis X-ray images, where features are extracted using the pre-trained ResNet18 model.
Keywords:
Support Vector Machine
Large-scale classification
Linex loss function
Journal
C
IF:
1.1
Papers:
25
Citations:
0
