灰色系统残差与Verhulst模型预测恶性肿瘤死亡  

Application of the Grey System Residual Error Model and Grey Verhulst Model on Forecasting Malignant Tumor Death

作　　者：孙彬[1] 何穗智[1] 

机构地区：[1]中山大学公共卫生学院预防医学研究所医学统计与流行病学系,广东广州510080

出　　处：《预防医学情报杂志》2011年第8期604-607,共4页Journal of Preventive Medicine Information

摘　　要：目的研究恶性肿瘤的流行趋势及其预测,为恶性肿瘤的综合防治提供准确、可靠、科学的依据。方法 根据香港特别行政区居民2001-2009年恶性肿瘤死亡资料建立灰色系统残差GM(1,1)模型与Verhulst模型,预测2001-2009年该市居民恶性肿瘤流行趋势。结果恶性肿瘤病死率趋势残差GM(1,1)模型精度检验结果显示:C=0.29,P=1.00,平均相对误差为1.01%,模型判为优。应用残差GM(1,1)预测2010-2013年恶性肿瘤病死率分别为183.2/10万、184.6/10万、186.1/10万和187.6/10万;用相邻做差的方式建立2002-2009年每年新增病死人数序列,每年新增病死人数的灰色Verhulst模型精度检验结果显示:C=0.634 2,P=0.45,平均相对误差为67.88%,模型结果提示效果较差。结论 分析结果说明香港特别行政区居民恶性肿瘤死亡率呈持续增加趋势,同时Verhulst模型结果提示香港每年新增病死人数尚未达到饱和状态,即未来几年每年新增恶性肿瘤死亡人数可能会有较大增幅。Objective To investigate the epidemic trend and its prediction,provide accurate,reliable scientific basis for comprehensive prevention of malignant tumor.Methods According to the malignant tumor death data of residents of Hong Kong SAR during 2001-2009,the grey system residual model GM（1,1） and Verhulst model were established,the epidemic trend of malignant tumor among city residents during 2001-2009 was predicted.Results Malignant tumor mortality trends residual GM（1,1） model accuracy testing results showed C=0.29 and P=1.00,the average relative error was 1.01%,and the model was optimal.Applicated residual GM（1,1） to predict the malignant tumor fatality rate during 2010-2013,the results were 183.2/105,184.6/105,186.1/105 and 187.6/105 respectively.Annually newly increased die amount series during 2002-2009 was created by using adjacent subtraction way,the Grey Verhulst model accuracy testing results showed： C=0.6342 and P=0.45,the average relative error was 67.88%,model results suggested that effect was relatively poor.Conclusion Results show that the malignant tumor mortality among residents of the Hong Kong SAR is increasing,while Verhulst model results suggest that Hong Kong annually die number has not reached saturation yet,that means the annually malignant tumor death amount may have substantial increase in the coming years.

关 键 词：灰色系统残差模型 灰色VERHULST模型 恶性肿瘤 预测 

分 类 号：R73[医药卫生—肿瘤]