返回
A particle swarm optimization-based feature selection for unsupervised transfer learning
DOI:10.1007/s00500-020-05105-1.png)
摘要
En 中文
Transfer learning (TL) method has captured an attractive presence because it facilitates the learning ability in the target domain by acquiring knowledge from well-established source domains. To gain strong knowledge from the source domain, it is important to narrow down the distribution difference between the source and the target domains. For this purpose, it is necessary to consider the objectives such as preserving the discriminative information, preserving the original similarity of the source and the target domain data, maximizing the variance of the target domain, and preserving marginal and conditional distribution at the same time. Furthermore, some existing TL methods use only original feature data, so there is a threat of degenerated feature transformation. To overcome all these limitations, in this paper, a novel feature selection-based transfer learning approach using particle swarm optimization (PSO) for unsupervised transfer learning (FSUTL-PSO) is implemented. In FSUTL-PSO, we incorporate all such objectives into one fitness function and select common good features from the source and target domains based on the fitness function for eliminating the threat of degenerated features. Extensive experiments have been done on all possible tasks of Office+Caltech and PIE Face datasets and our proposed method FSUTL-PSO has shown significant improvement over the existing transfer or non-transfer learning methods.
Keyword:
Domain adaptation
Classification
Particle swarm optimization
Transfer learning
Unsupervised discriminant analysis
Feature selection
AI总结
对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
2.5
论文数:
1.0W
被引数:
2.1W
机构
引用论文
Autologous graft-versus-host disease: harnessing anti-tumor immunity through impaired self-tolerance
Association of Single-Nucleotide Polymorphisms in Immune-Related Genes with Development of Dengue Hemorrhagic Fever in a Mexican Population免疫相关基因单核苷酸多态性与墨西哥人群登革出血热发生发展的关联
Evolution in sown populations of subterranean clover (Trifolium subterraneum L. ) in South Australia

