arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.03498cs.AI
收藏

面向次日出院预测的资源感知建模

Resource-Conscious Modeling for Next- Day Discharge Prediction Using Clinical Notes

University of California, Irvine, Donald Bren School of Information, Department of Informatics(加州大学尔湾分校,唐纳德·布伦信息与计算机科学学院,信息学系);University of California, Irvine, Department of Neurosurgery(加州大学尔湾分校,神经外科系)

Ha Na Cho, Sairam Sutari, Alexander Lopez, Hansen Bow, Kai Zheng

AI总结:

本文研究了轻量级微调大语言模型和传统文本模型在次日出院预测中的应用,发现TF-IDF与LGBM在平衡性上表现最佳,AUC-ROC达0.80,表明可解释性模型在临床任务中更具优势。

摘要:

及时的出院预测对于优化 elective spine surgery 单位的床周转和资源分配至关重要。本研究评估了轻量级微调大语言模型(LLMs)和传统文本模型在使用术后临床笔记进行次日出院预测的可行性。我们比较了13种模型,包括TF-IDF与XGBoost、LGBM以及通过LoRA微调的紧凑型LLMs(DistilGPT-2、Bio_ClinicalBERT)。TF-IDF与LGBM在平衡性上表现最佳,出院类别的F1得分为0.47,召回率为0.51,AUC-ROC最高为0.80。尽管LoRA提高了DistilGPT2的召回率,但整体上基于Transformer和生成模型的表现不佳。这些发现表明,在实际的不平衡临床预测任务中,可解释性且资源高效的模型可能优于紧凑型LLMs。

英文摘要:

Timely discharge prediction is essential for optimizing bed turnover and resource allocation in elective spine surgery units. This study evaluates the feasibility of lightweight, fine-tuned large language models (LLMs) and traditional text-based models for predicting next-day discharge using postoperative clinical notes. We compared 13 models, including TF-IDF with XGBoost and LGBM, and compact LLMs (DistilGPT-2, Bio_ClinicalBERT) fine-tuned via LoRA. TF-IDF with LGBM achieved the best balance, with an F1-score of 0.47 for the discharge class, a recall of 0.51, and the highest AUC-ROC (0.80). While LoRA improved recall in DistilGPT2, overall transformer-based and generative models underperformed. These findings suggest interpretable, resource-efficient models may outperform compact LLMs in real-world, imbalanced clinical prediction tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑