运用ARIMA模型预测江岸区恶性肿瘤死亡率的变化趋势  

Application of ARIMA model in predicting the cancer mortality in Jiang＇an district, Wuhan

作　　者：朱慈华 丁彦培 叶波[2] 冯仁杰[2] 马露[2] 

机构地区：[1]武汉市江岸区疾病预防控制中心,湖北武汉430017 [2]武汉大学公共卫生学院,湖北武汉430017

出　　处：《中华临床医师杂志（电子版）》2016年第7期41-42,共2页Chinese Journal of Clinicians(Electronic Edition)

摘　　要：目的 了解2003-2013 年江岸区恶性肿瘤死亡现况及其变化趋势,预测江岸区2014 年的恶性肿瘤死亡率,探讨ARIMA 模型（自回归移动平均模型）在恶性肿瘤死亡预测中的应用.方法 运用时间序列分析法对江岸区2003-2013 年恶性肿瘤死亡数据进行分析并建立预测模型,通过模型识别、参数估计与模型诊断选择最优预测模型.结果 2003-2013 年江岸区恶性肿瘤年均死亡率为174.37/10 万,且存在上升趋势.建立模型ARIMA（0,1,1）（0,1,1）12,其拟合度和预测效果较为理想.结论 ARIMA 模型能较好的模拟江岸区恶性肿瘤的死亡趋势,可用于短期预测和分析.Objective Applying ARIMA model to explore the cancer mortality trend in Jiang＇an district from 2003 to 2013 and predict numbersof deaths from cancer up to 2014. Methods In this research, we used a time series analysis of cancer death rates from 2003 to 2013 in Jiang＇an districtand established ARIMA model based on model identification, estimation and verification of parameter, and analysis of the fitting of model. Results From2003 to 2013 in Jiang＇an district, there was an overall upward cancer morality trend and the annual average of cancer mortality was 174.37 per 100 000.Besides, ARIMA （0,1,1） （0,1,1）12 model can be used as the optimal model due to its high forecasting accuracy. Conclusion ARIMA （0,1,1） （0,1,1）12model can be employed to simulate cancer mortality trend in Jiang＇an district and it is a predicted model of high precision for short term forecast.

关 键 词：ARIMA模型 时间序列分析 恶性肿瘤 死亡率 

分 类 号：R195.1[医药卫生—卫生统计学]