软标签恢复基于标签特定特征的学习学习
Jiansheng Jiang1, Wenxin Ge2, Yibin Wang3
1School of Computer and Information, Anqing Normal University, Anqing, 246133, China.
Scientific reports
|October 4, 2024
概括
这项研究引入了多标签分类的新算法,该算法解决了缺失标签和标签错误分类的问题. 基于软标签恢复的标签特定特征学习 (SLR-LSF) 方法通过创建更丰富的软标签来提高分类准确性.
科学领域:
- 机器学习 机器学习
- 数据挖掘 数据挖掘
- 计算机科学 计算机科学
背景情况:
- 多标签分类通常使用二进制逻辑标签,导致错误分类和数据集中缺少标签的问题.
- 现有的算法通常只解决其中一个挑战 (缺失或错误分类的标签),因此需要采用更全面的方法.
研究的目的:
- 为了提出一个新的算法,软标签恢复基于标签特定特征学习 (SLR-LSF),能够同时解决标签错误分类和多标签数据集中缺失的标签.
- 开发一种构建软标签的方法,可以准确地反映实例-标签关系,并包含更丰富的语义信息.
主要方法:
- 利用信息来计算标签之间的信任矩阵.
- 将标签密度信息与会员级别结合起来,以构建软标签,有效处理缺失的标签.
- 采用流规范化和全球标签相关性来学习标签特定特征,增强本地流性和整体分类性能.
主要成果:
- 拟议的SLR-LSF算法成功地恢复了缺失的标签,并生成了带有增强语义信息的软标签.
- 在多个数据集上的实验结果表明,与现有方法相比,SLR-LSF在提高多标签分类性能方面的有效性.
- 在特征学习过程中,将本地平滑度和全球标签相关性整合到特征学习过程中,有助于优异的分类结果.
结论:
- SLR-LSF算法提供了一个统一的框架,用于解决标签错误分类和多标签分类中缺失的标签.
- 开发的软标签结构和标签特定特征学习机制显著提高了分类准确性.
- 这项研究为提高多标签分类系统的可靠性和性能提供了强大的解决方案.
更多相关视频
07:44Label-Retention Expansion Microscopy LR-ExM Enables Super-Resolution Imaging and High-Efficiency Labeling
Published on: October 11, 2022
3.6K
10:36High Resolution Quantitative Synaptic Proteome Profiling of Mouse Brain Regions After Auditory Discrimination Learning
Published on: December 15, 2016
10.5K
相关概念视频
01:20Labeling Emotion
114
Emotional labeling is a cognitive process that involves identifying and naming one's emotions, such as anger, fear, happiness, or sadness. It allows individuals to recognize and express their internal emotional states, a critical aspect of emotional regulation and communication. Labeling emotions requires more than mere recognition; it also involves drawing upon memory and contextual cues to understand the current situation and apply a corresponding emotional label. For instance, feeling...
114
03:31Labeling DNA Probes
8.1K
DNA probes are fragments of DNA labeled with a reporter tag to enable their detection or purification. The resulting labeled DNA probes can then hybridize to target nucleic acid sequences through complementary base-pairing, and may be used to recover or identify these regions.
Radioisotopes, fluorophores, or small molecule binding partners like biotin or digoxigenin, are the most widely used reporter tags for labeling DNA probes. These labels can be attached to the probe DNA molecule via...
Radioisotopes, fluorophores, or small molecule binding partners like biotin or digoxigenin, are the most widely used reporter tags for labeling DNA probes. These labels can be attached to the probe DNA molecule via...
8.1K
01:23Principles of Classical Conditioning
511
Classical conditioning, as described by Ivan Pavlov, is a foundational concept in associative learning, where a neutral stimulus becomes capable of eliciting a conditioned response through association with an unconditioned stimulus. The process of acquisition, where this learning occurs, and the subsequent phenomena of contiguity, contingency, generalization, discrimination, extinction, and spontaneous recovery are crucial for a comprehensive understanding of classical conditioning.
During the...
During the...
511