ARIMA乘积季节模型与广义回归神经网络模型在布鲁菌病发病预测的比较

doi:10.6040/j.issn.1671-7554.0.2017.1185

摘要/Abstract

摘要： 目的探讨适合全国布鲁菌病发病的预测模型,为布鲁菌病预测预警系统提供参考。方法利用中国疾病预防控制中心2011年1月至2016年12月按月报告的布鲁菌病发病数历史疫情数据,分别建立求和(差分)自回归移动平均(ARIMA)乘积季节预测模型和广义回归神经网络(GRNN)模型,对2017年1~8月月报数进行预测,采用实际发病数与两种模型预测数进行比较,评价指标为平均相对误差、平均绝对误差。结果建立的ARIMA(0,1,1)(0,1,1)₁₂乘积季节模型平均绝对误差、平均相对误差分别是989、0.23,GRNN模型平均绝对误差、平均相对误差分别是561、0.14,均小于ARIMA模型。结论 ARIMA模型和GRNN模型均可用于预测布病的发病数,后者预测效能优于前者。

关键词: 求和(差分)自回归移动平均, 乘积季节模型, 布鲁菌病, 广义回归神经网络模型, 时间序列, 预测

Abstract: Objective To explore suitable model for brucellosis incidence forecasting in China, and to provide reference for forecasting warning system of brucellosis. Methods Autoregressive integrated moving average(ARIMA)model and generalized regression neural network(GRNN)model were fitted with data monthly reported by China Centers for Disease Control from January 2011 to December 2016. The monthly reported data from January to August 2017 were used to evaluate forecast results. The mean absolute error(MAE)and mean relative error(MRE)were evaluated by comparing the actual incidence with the predicted incidence of the two models. Results The MAE and MRE of the ARIMA(0,1,1)(0,1,1)₁₂model were 989, 0.23 and the GRNN model were 561, 0.14, respectively. The MAE and MRE of the GRNN model were less than the ARIMA model. Conclusion Both the ARIMA and GRNN model perform well in forecasting the incidence of brucellosis, while the prediction ability of GRNN model is slightly better than ARIMA model.

Key words: Autoregressive integrated moving average, Multiple seasonal model, Brucellosis, Generalized regression neural network model, Time series, Forecast

中图分类号:

R183

马洁,田野,黄璐,孟维静,王素珍,石福艳. ARIMA乘积季节模型与广义回归神经网络模型在布鲁菌病发病预测的比较[J]. 山东大学学报 (医学版), 2018, 56(6): 76-82.

MA Jie, TIAN Ye, HUANG Lu, MENG Weijing, WANG Suzhen, SHI Fuyan. Comparison of multiple seasonal ARIMA model and generalized regression neural network model in forecasting the incidence of brucellosis[J]. Journal of Shandong University (Health Sciences), 2018, 56(6): 76-82.

参考文献

[1] 田德红, 于国伟, 丁国武, 等. ARIMA-DES混合模型在中国布鲁菌病分析和预测中的应用[J]. 中国卫生统计, 2016, 33(2): 245-248. TIAN Dehong, YU Guowei, DING Guowu, et al. Application of ARIMA-DES hybrid model in the analysis and prediction of Chinese brucellosis[J]. Chinese Journal of Health Statistics, 2016, 33(2): 245-248.
[2] 娄鹏威, 王娜, 罗冬梅, 等. 基于季节性动力学模型预测与控制新疆布鲁氏菌病的流行[J]. 公共卫生与预防医学, 2017, 28(1): 13-17. LOU Pengwei, WANG Na, LUO Dongmei, et al. Forecasting and controlling the epidemic of brucellosis in Xinjiang based on seasonal dynamics model[J]. Public Health and Preventive Medicine, 2017, 28(1): 13-17.
[3] 李志军, 范凯, 宋颖. 辽宁省凌海市2011—2015年布病的发病症状及暴露危险因素调查[J]. 锦州医科大学学报, 2017, 38(2): 85-87. LI Zhijun, FAN Kai, SONG Ying. Investigation on the symptoms and exposure risk factors of the disease in Linghai city, Liaoning province from 2011 to 2015[J]. Journal of Jinzhou Medical University, 2017, 38(2): 85-87.
[4] 樊兆峰, 马小平, 邵晓根. 非线性系统RBF神经网络多步预测控制[J]. 控制与决策, 2014, 29(7): 1274-1278. FAN Zhaofeng, MA Xiaoping, SHAO Xiaogen. Multi-step predictive control of nonlinear system RBF neural network[J].Control and Decision, 2014, 29(7): 1274-1278.
[5] 戴文战, 娄海川, 杨爱萍. 非线性系统神经网络预测控制研究进展[J]. 控制理论与应用, 2009, 26(5): 521-530. DAI Wenzhan, LOU Haichuan, YANG Aiping. An overview of neural network predictive control for nonlinear systems[J]. Control Theory & Applications, 2009, 26(5): 521-530.
[6] Goh BS. Convergence of algorithms in optimization and solutions of nonlinear equations[J]. J Optim Theory Appl, 2010, 144(1): 43-55.
[7] 张文茜, 苏海霞, 尚磊, 等. 基于BP神经网络和RBF神经网络预测老年痴呆症疾病进展的对比研究[J]. 现代生物医学进展, 2017, 17(4): 738-741. ZHANG Wenqian, SU Haixia, SHANG Lei, et al. A comparative study on the progression of alzheimers disease based on BP neural network and RBF neural network[J]. Advances in modern biomedical science, 2017, 17(4): 738-741.
[8] 王永斌, 许春杰, 尹素凤, 等. 中国手足口病发病率ARIMA、RBF及ARIMA-RBF组合模型拟合及预测效果比较[J]. 中国公共卫生, 2017, 33(5): 760-763. WANG Yongbin, XU Chunjie, YIN Sufeng, et al. The combined models of ARIMA, RBF and ARIMA-RBF were compared[J]. Chinese Journal of Public Health, 2017, 33(5): 760-763.
[9] 刘红杨, 刘洪庆, 李望晨, 等. 差分自回归移动平均与广义回归神经网络组合模型在丙型肝炎月发病率中的预测应用[J].中国全科医学, 2017, 20(2): 182-186. LIU Hongyang, LIU Hongqing, LI Wangchen, et al. The predictive application of differential self-regression moving average and generalized regression neural network model in the incidence of hepatitis C[J]. Chinese general medicine, 2017, 20(2): 182-186.
[10] 孙振球.医学统计[M]. 北京: 人民出版社, 2010: 66-70.
[11] 沈冰, 高洁, 胡宏, 等. 上海静安区儿童流感样病例就诊百分比预测的自回归求和移动平均模型构建与应用[J]. 公共卫生与预防医学, 2017, 28(4): 58-61. SHEN Bing, GAO Jie, HU Hong, et al. The construction and application of the self-regression summation moving average model for the percentage prediction of pediatric influenza-like cases in Shanghai Jingan district[J]. Public health and Preventive medicine, 2017, 28(4):58-61.
[12] 朱蒙曼, 杨其松, 谢昀, 等. 江西省肾综合征出血热发病率ARIMA模型及其趋势预测[J]. 现代预防医学, 2017, 44(22): 4036-4041. ZHU Mengman, YANG Qisong, XIE Yun, et al. ARIMA model and its trend prediction of hemorrhagic fever with renal syndrome in Jiangxi province[J]. Modern Preventive Medicine, 2017, 44(22): 4036-4041.
[13] 王燕. 应用时间序列分析[M]. 3版. 北京: 中国人民大学出版社, 2012: 158-159.
[14] 梅树江, 周志峰, 马汉武, 等. 深圳市ARIMA在肾综合征出血热发病预测中应用[J]. 中国公共卫生, 2015, 31(7): 936-938. MEI Shujiang, ZHOU Zhifeng, MA Hanwu, et al. The application of ARIMA in the prediction of hemorrhagic fever in renal syndrome[J]. Chinese Journal of Public Health, 2015, 31(7): 936-938.
[15] 梁纪伟, 姜法春, 韩雅琳,等. 应用ARIMA乘积季节模型预测青岛市甲肝发病[J]. 中国公共卫生管理, 2016, 32(6): 780-782,793. LIANG Jiwei, JIANG Fachun, HAN Yalin, et al. The multiple seasonal ARIMA model was used to predict the incidence of hepatitis a in Qingdao[J]. Chinese Journal of Public Health Management, 2016, 32(6): 780-782,793.
[16] 朱奕奕, 冯玮, 赵琦, 等. ARIMA乘积季节模型在上海市甲肝发病预测中的应用[J]. 复旦学报(医学版), 2012, 39(5): 460-464. ZHU Yiyi, FENG Wei, ZHAO Qi, et al. The application of multiple seasonal ARIMA model in the prediction of hepatitis A in Shanghai[J]. Journal of Fudan university(medical edition), 2012, 39(5): 460-464.
[17] Donnan EJ, Fielding JE, Gregory JE, et al. A multistate outbreak of hepatitis a associated with semidried tomatoes in Australia, 2009[J]. Clin Infect Dis, 2012, 54(6): 775-781.
[18] 王文霞,王春红,葛少磊.基于广义回归神经网络的图像修复算法[J].计算机工程与设计,2017,38(11):3125-3130. WANG Wenxia, WANG Chunhong, GE Shaolei. Image restoration algorithm based on generalized regression neural network[J]. Computer engineering and design, 2017, 38(11):3125-3130.
[19] Wang Y, Nie N, Wang M, et al. Mine tailings facilities safety evaluation of GRNN optimized by modified fruit fly algorithm[J]. Computer Engineering, 2015, 41(4): 267-272.
[20] 江帆, 刘辉, 王彬, 等. 基于CNN-GRNN模型的图像识别[J]. 计算机工程, 2017, 43(4): 257-262. JIANG Fan, LIU Hui, WANG Bin, et al. Image recognition based on CNN-GRNN model[J]. Computer Engineering, 2017, 43(4): 257-262.
[21] 卢金娜. 基于优化算法的径向基神经网络模型的改进及应用[D]. 太原: 中北大学, 2015.
[22] 杨召, 叶中辉, 尤爱国,等. 乘积季节ARIMA模型在结核病发病率预测中应用[J]. 中国公共卫生, 2013, 29(4): 469-472. YANG Zhao, YE Zhonghui, YOU Aiguo, et al. Multiple seasonal ARIMA model was applied in the prediction of tuberculosis incidence[J]. Public health in China, 2013, 29(4): 469-472.
[23] Wu W, Guo J, An S, et al. Comparison of two hybrid models for forecasting the incidence of hemorrhagic fever with renal syndrome in Jiangsu Province, China[J]. PLoS One, 2015, 10(8): e0135492. doi:10.1371/journal.
[24] 魏仁惠子, 沈双全, 欧春泉. SARIMA模型与SARIMA-GRNN组合模型在预测广东省登革热疫情中的应用[J]. 中国卫生统计, 2016, 33(5): 746-748. WEI Renhuizi, SHEN Shuangquan, OU Chunquan. The SARIMA model and the SARIMA-GRNN model are used to predict dengue fever in Guangdong province[J]. Chinese Journal of Health Statistics, 2016, 33(5):746-748.
[25] 孟凡东, 吴迪, 隋承光. 2004—2015年中国狂犬病发病数据ARIMA乘积季节模型的建立及预测[J]. 中国卫生统计, 2016, 33(3): 389-391, 395. MENG Fandong, WU Di, SUI Chengguang. The establishment and prediction of the multiple seasonal ARIMA model of rabies in China from 2004 to 2015[J]. Chinese Journal of Health Statistics, 2016, 33(3): 389-391,395.

相关文章 15

[1]	王烨,王刚,南海鸥,蔡雨玲,李嘉宇,王海锋,颜英杰,江志伟. 胃癌患者围手术期心率变异度与炎性指标的相关性[J]. 山东大学学报 (医学版), 2026, 64(4): 23-30.
[2]	段盈竹,董波,于睿. 内在情感与类风湿关节炎患者冠状动脉粥样硬化风险关系的孟德尔随机化分析[J]. 山东大学学报 (医学版), 2026, 64(4): 63-71.
[3]	邹宇锦,万熠,纪振冰,梁西昌. 促骨再生-抗肿瘤双功能钛合金植入体的生物实验研究[J]. 山东大学学报 (医学版), 2026, 64(2): 50-65.
[4]	王建民,李晓峰,由志涛,董圣杰,赵宇驰,李占菊,邹德鑫,张剑锋,孙涛,杜伟. 基于可解释机器学习的后路腰椎椎体间融合术后慢性疼痛风险预测模型构建[J]. 山东大学学报 (医学版), 2026, 64(2): 78-88.
[5]	孙爽爽,仉率杰,张伯韬,袁莹,于媛媛,薛付忠. 基于真实世界研究的18~50岁人群急性缺血性卒中影响因素[J]. 山东大学学报 (医学版), 2025, 63(9): 40-46.
[6]	王梦星,薛付忠,杨帆. 基于多模态交叉注意力机制融合的1型糖尿病血糖浓度预测方法[J]. 山东大学学报 (医学版), 2025, 63(8): 41-50.
[7]	申路佳,逯天威,巩伟明,赵岩松,王淑康,袁中尚. 代谢风险评分在2型糖尿病人群心血管结局预测中的应用[J]. 山东大学学报 (医学版), 2025, 63(8): 69-78.
[8]	李千,杨帆,薛付忠. 基于多模态数据融合的多癌种风险预测模型[J]. 山东大学学报 (医学版), 2025, 63(8): 79-85.
[9]	李晓琪,刘佩丽,成红,赵艳艳. 基于自注意力机制预测ICU脓毒症患者的死亡率[J]. 山东大学学报 (医学版), 2025, 63(8): 86-93.
[10]	陈莹莹,王鲁,胡锡峰,朱高培,薛付忠. 基于贝叶斯网络的2型糖尿病患者并发脑卒中风险预测[J]. 山东大学学报 (医学版), 2025, 63(8): 94-102.
[11]	王丽云,高天勤,刘雨佳,陈青,陈柳,沙凯辉. 基于机器学习产后压力性尿失禁风险预测模型的构建及验证[J]. 山东大学学报 (医学版), 2025, 63(6): 55-66.
[12]	杜雪,李春霞,刘云霞,张涛. 基于MFPC-Cox的结直肠癌患者预后动态预测模型[J]. 山东大学学报 (医学版), 2025, 63(5): 101-110.
[13]	陈瑛翼,游倩,王意,张帆,李凤,季舒铭,徐浩源,饶志勇. 办公室职员肌肉质量减少预测模型的开发与验证[J]. 山东大学学报 (医学版), 2025, 63(4): 26-35.
[14]	王宝炫,焦杰,张厚君,刘奇,于冠英. 衰弱与肌少症评估在胃肠道肿瘤术后结局预测中的应用与展望[J]. 山东大学学报 (医学版), 2025, 63(4): 51-58.
[15]	张骁驰, 吕婷婷, 于文浩, 李国傲, 高杉杉, 赵琦, 王立友. 2014—2019年我国极端降水与其他感染性腹泻的关联性研究[J]. 山东大学学报 (医学版), 2025, 63(4): 100-105.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed