计算机科学
医疗保健
医疗保健系统
机器学习
风险分析(工程)
人工智能
癌症
医疗保健政策
心理干预
预测建模
数据挖掘
计算复杂性理论
计算模型
变压器
癌症治疗
作者
Xing Liu,Byung‐Gyu Kim,Jianhui Lv,Zhuo Jia Wu
标识
DOI:10.1109/jbhi.2025.3622630
摘要
This paper presents BiSAUM, a bi-directional sparse attention transformer framework that simultaneously analyzes sustainable healthcare patterns and cancer cell characteristics while maintaining high computational efficiency. The model introduces a novel bi-directional sparse selection mechanism that reduces computational complexity while preserving crucial relationships between healthcare delivery systems and cancer cell behaviors. Through experiments, BiSAUM demonstrates superior performance in both sustainable healthcare prediction and cancer cell pattern classification, particularly in understanding how healthcare interventions affect cancer progression. The model achieves 12.1% lower MSE in healthcare system prediction and 4.4% higher accuracy in cancer pattern classification compared to state-of-the-art baselines while reducing computational time by 35.8% and memory usage by 32.6%. These results demonstrate BiSAUM's potential for advancing our understanding of how multi-domain sustainable healthcare systems impact cancer progression, enabling more effective, evidence-based medical decisions in cancer prevention and treatment.
科研通智能强力驱动
Strongly Powered by AbleSci AI