陆宇，江会

引用本文： 陆宇，江会.基于3种机器学习算法构建宫颈癌术后尿潴留风险预测模型[J].护理研究,2024,38(1):24-30

摘要： 目的： 运用决策树、逻辑回归和支持向量机构建宫颈癌根治性切除术后尿潴留风险预测模型并比较性能，为评估及预防宫颈癌术后尿潴留提供参考依据。 方法： 回顾性收集459例宫颈癌根治性切除术病人的临床资料，采用决策树、支持向量机和逻辑回归3种机器学习方法构建宫颈癌根治性切除术后尿潴留风险预测模型，采用准确性、召回率、精确率、F1指数和受试者工作特征（ROC）曲线下面积（AUC）评价模型性能。 结果： 共纳入病人的年龄、疾病分期、体质指数等8个变量。选择80%的数据集（367例）作为训练集，20%的数据集（92例）作为验证集，结果显示，决策树在训练集和验证集中准确率、召回率、精确率、F1指数和AUC都比支持向量机和逻辑回归更优，说明决策树在构建宫颈癌术后尿潴留风险预测模型中具有较高的准确率及较好的泛化性能；支持向量机在训练集中准确率、召回率、精确率、F1指数和AUC都比逻辑回归更优。同时，在验证集中，支持向量机的召回率和F1指数比逻辑回归更优，但是支持向量机的准确率、精确率和AUC却比逻辑回归差，说明支持向量机在宫颈癌术后尿潴留数据集中的泛化能力比逻辑回归差。 结论： 决策树在构建宫颈癌根治性切除术后尿潴留风险预测模型中具有较高的性能及较好的泛化能力，可为相关临床决策提供指导建议。

关键词： 宫颈癌;尿潴留;危险因素;机器学习;预测模型;决策树;支持向量机;逻辑回归

课题： 上海申康医院发展中心管理研究项目(2022SKMR⁃18)

LU Yu，JIANG Hui

Abstract Objective： To establish risk prediction model of postoperative urinary retention using decision tree，logistic regression and support vector machine and compare their performance，in order to provide references for evaluating and preventing urinary retention after radical resection. Methods： The medical history information of 459 patients who underwent radical resection in Shanghai First Maternal and Infant Health Hospital from 2018 to 2021 was collected retrospectively.Three machine learning models，decision tree，logistic regression and support vector machine，were used to construct the risk prediction model of postoperative urinary retention.Accuracy，recall，precision，F1 score and AUC were used to evaluate the performance of the models. Results： A total of eight variables were included， including age， disease stage， body mass index and so on.80% of data sets （367 cases） were selected as training sets and 20%（92 cases） were selected as test sets.The results showed that the accuracy，recall，precision， F1 and AUC of decision tree in the training set and test set were better than those of support vector machine and logistic regression，which indicated that decision tree had higher accuracy and better generalization performance. SVM was better than logistic regression in accuracy， recall， precision， F1 and AUC in the training set. At the same time， in the test sets， the recall and F1 of support vector machine were better than those of logistic regression； but the accuracy， precision and AUC of support vector machine were worse than those of logistic regression， indicating that the generalization ability of support vector machine was worse than that of logistic regression. Conclusion： Decision tree has high performance and generalization ability in constructing the risk prediction model of postoperative urinary retention， which can provide guidance and suggestions for related clinical behaviors.

Keywords cervical cancer;urinary retention;risk factors;machine learning;prediction model;decision tree;support vector machine;Logistic regression

阅读全文：0

免费浏览,查看全文　 html版　 　PDF版
阅读此文章需pdf阅读器 下载Pdf格式阅读器