基于ARIMA与NNAR模型的甘肃省胰腺癌发病趋势预测  

Prediction of pancreatic cancer incidence trend in Gansu province based on ARIMA and NNAR model

作　　者：马晨哲 王霄 杨波 郭继武[1,2] 李玉民 MA Chenzhe;WANG Xiao;YANG Bo;GUO Jiwu;LI Yumin(The Second Clinical Medical College,Lanzhou University,Gansu Lanzhou 730030,China;Key Laboratory of Digestive System Tumors of Gansu Province,Gansu Lanzhou 730030,China)

机构地区：[1]兰州大学第二临床医学院,甘肃兰州730030 [2]甘肃省消化系统肿瘤重点实验室,甘肃兰州730030

出　　处：《现代肿瘤医学》2024年第11期2080-2084,共5页Journal of Modern Oncology

基　　金：甘肃省教育厅项目(编号:2021jyjbgs-02);甘肃省发改委项目(编号:2020-2022);甘肃省科技重大专项(编号:20ZD7FA003);甘肃省军民融合发展专项(编号:2020-128)。

摘　　要：目的:分析2013-2021年甘肃省胰腺癌发病趋势,并对2022-2026年胰腺癌发病趋势进行预测。方法:收集2013-2021年甘肃省胰腺癌发病数据,通过计算年度变化百分比(annual percentage change,APC)和平均年度变化百分比(average annual percentage change,AAPC)分析其变化趋势;构建自回归移动平均(autoregressive integrated moving average,ARIMA)模型和神经网络自回归(neural network autoregression,NNAR)模型,比较两种模型预测精度并预测2022-2026年胰腺癌发病趋势。结果:选择精度更佳的ARIMA模型预测得到2022-2026年甘肃省胰腺癌发病率分别为5.58/10万、6.01/10万、6.44/10万、6.88/10万、7.31/10万。甘肃省胰腺癌发病率在2022-2026年仍保持上升趋势。结论:2022-2026年甘肃省胰腺癌发病率呈现上升趋势,ARIMA模型预测发病趋势具有良好的精度,可以为进一步疾病防控工作提供重要参考依据。Objective:To evaluate the pancreatic cancer incidence trend in Gansu Province from 2013 to 2021 and forecast the pancreatic cancer incidence trend from 2022 to 2026.Methods:Gather statistics on pancreatic cancer incidence in Gansu Province between 2013 and 2021,and analyze the trend of change by calculating APC and AAPC.Establish ARIMA model and NNAR model,compare the prediction accuracy of the two models and predict the incidence trend of pancreatic cancer from 2022 to 2026.Results:The ARIMA model with better accuracy is used to predict the incidence of pancreatic cancer in Gansu Province from 2022 to 2026 to be 5.58/100000,6.01/100000,6.44/100000,6.88/100000,and 7.31/100000,respectively.The incidence of pancreatic cancer in Gansu Province will still maintain an upward trend from 2022 to 2026.Conclusion:The incidence of pancreatic cancer in Gansu Province will show an upward trend from 2022 to 2026.The ARIMA model predicts the incidence trend with adequate precision,which can provide an important reference for further disease prevention and control.

关 键 词：胰腺癌 发病趋势 预测 ARIMA模型 NNAR模型 

分 类 号：R735.9[医药卫生—肿瘤]