基于GA-TabNet-CatBoost模型的乳腺癌患者住院费用预测方法OA
A method for predicting hospitalization costs of breast cancer patients based on the GA-TabNet-CatBoost model
结合患者的历史数据、入院前数据及住院期间数据,提出了一种基于 GA-TabNet-CatBoost 模型的乳腺癌患者住院费用预测方法.该方法采用两阶段设计,首先通过 GA-TabNet 模型预测患者的住院时长,然后结合住院期间的数据通过 CatBoost 模型预测患者的住院费用.结果表明,GA-TabNet-CatBoost 模型预测方法在乳腺癌患者住院费用预测中的R2 值达 0.889.该方法可在无离院先验信息的条件下,完成住院时长与住院费用预测,能为临床医疗决策与医疗成本管控提供更灵活的解决方案.
Combining patients'previous records,pre-admission information,and in-hospital data,this paper proposed a method for predicting hospitalization costs of breast cancer patients based on the GA-TabNet-CatBoost(GTC)model.The method adopted a two-stage design.First,the GA-TabNet model was used to predict the length of hospital stay.Then the CatBoost model was combined with in-hospital data to predict hospitalization costs.The results showed that the GTC model achieved an R2 value of 0.889 in predicting hospitalization costs for breast cancer patients.This method can simultaneously predict both length of stay and hospitalization costs without requiring prior discharge information,providing a more flexible solution for clinical medical decision-making and healthcare cost management.
朱海玉;杨丽静;孙晓燕;袁贞明
杭州师范大学信息科学与技术学院,浙江 杭州||商丘职业技术学院轻工业学院,河南 商丘杭州市卫生健康事业发展中心,浙江 杭州杭州师范大学信息科学与技术学院,浙江 杭州杭州师范大学信息科学与技术学院,浙江 杭州
信息技术与安全科学
住院费用预测TabNetCatBoost乳腺癌
hospitalization cost predictionTabNetCatBoostbreast cancer
《杭州师范大学学报(自然科学版)》 2026 (5)
468-475,8
杭州市卫生科技计划重点基金项目(ZD20230018)
评论