对缩小维度和类平衡进行实证评估,用于医学文本分类
Arslan Jamil1, Muhammad Kashif Hanif2, Muhammad Umer Sarwar1
1Department of Computer Science, Government College University, Faisalabad, Pakistan.
Scientific reports
|December 12, 2025
概括
一个新的模型有效地使用主成分分析 (PCA) 和合成少数人过量采样技术 (SMOTE) 来压缩电子健康记录中的临床叙述. 这种方法提高了准确性,并大大加快了医疗保健中自然语言处理任务的培训.
科学领域:
- 医疗信息学 医疗信息学
- 自然语言处理自然语言处理.
- 机器学习 机器学习
背景情况:
- 电子健康记录包含了大量的非结构化的临床文本.
- 文本数据的高维度和稀疏性阻碍了准确和高效的分析.
- 临床叙述的自动分类对于数据驱动的医疗保健至关重要.
研究的目的:
- 开发一种轻量级模型来压缩纵向患者叙述.
- 提高临床文本分类的性能和计算效率.
- 为了使医疗叙述能够在不影响诊断准确性的情况下进行大规模分析.
主要方法:
- 使用多重学习技术开发了一种连续减小模型.
- 探讨了主要组件分析 (PCA),t分布式静态邻居嵌入 (t-SNE) 和统一的多重近似和投影 (UMAP).
- 这些技术与合成少数人过量采样技术 (SMOTE) 相结合.
主要成果:
- 在PCA和SMOTE的组合中,在MTSamples体上获得了91.2%的准确性.
- 在未降低的基线上观察到具有统计意义的6.4%的宏观F1增长.
- 模型训练时间减少了42%,提高了计算效率.
结论:
- PCA与SMOTE相结合,为临床NLP提供了一个实用且可扩展的解决方案.
- 这种方法可以提高计算效率,而不会牺牲诊断准确度.
- 这些发现支持在医疗机构部署高效的NLP管道来分析医疗叙事.
相关概念视频
01:21Dimensions of Health and Illness
10.1K
The factors influencing the health-illness continuum can be internal or external and may or may not be under conscious control. They are related to the following eight human dimensions, and each dimension is interrelated to one other.
10.1K
01:17Classification of Illness
8.5K
The meaning of illness is individualized to each person who experiences an alteration in health. In contrast, disease is a medical term indicating a pathological change in the structure and function of the body or mind. It is a condition that has specific symptoms and boundaries.
An illness is a response to a disease in which the person's level of functioning is changed compared with a previous level. The general classification of illness includes acute and chronic.
Acute illness is severe...
An illness is a response to a disease in which the person's level of functioning is changed compared with a previous level. The general classification of illness includes acute and chronic.
Acute illness is severe...
8.5K