江苏省启东市1972-2001年肺癌发病趋势分析及预测模型比较研究

Study on time-series analysis and forecast models on lung cancer incidence in Qidong, 1972-2001

目的 引入时间序列分析江苏省启东市1972-2001年肺癌发病资料并预测其到2010年的变化趋势。方法 利用启东市肿瘤登记处积累的肺癌1972-2001年的登记资料，采用时间序列分析方法中的趋势外推法、指数平滑法、自回归综合移动平均模型(ARIMA模型)等分析该地肺癌发病率的变化趋势。比较各方法预测精度，赋予不同权重，建立组合预测模型。结果 各种分析方法预测趋势结果：肺癌的发病率将呈现上升趋势；预测今后十年仍继续呈现上升趋势，估汁至2010年启东市男性肺癌发病率将突破70／10万，女性将达到20／10万，在2001年基础上将增长33％和10％以上，接近周边大城市肺癌发病率水平。结论 通过预测模型的分析和比较，认为运用时间序列分析预测肺癌发病率趋势是恰当的，建立组合模型可以提高预测精度。组合模型与ARIMA法预测启东市肺癌发病率结果相近，可以用ARIMA法作为肺癌发病率预测的主要方法。

Objective To explore the lung cancer incidence rates from 1972 to 2001 and utilize varieties of models in forecasting trend up to 2010 in the city of Qidong,Jiangsu in order to provide baseline data for its control and prevention. Methods Using data from the cancer registry office in Qidong, we tried to reveal the trends of lung cancer incidence by analyzing the time-series on trends extrapolation, exponent smoothness,Box-Jenkins model etc. We also compared the prognostication precision, endow differ power, and established assembled forecast model. Results Data showed that there had been a rising trend of lung cancer from 1972 to 2001 and would still probably be on the increase in the future. The rate of male and female attained to 70 per 100 000 and 20 per 100 000, predicting that there would be a respective 33 percent and 10 percent increase in 2010. Conclusions According to analysis of forecast models,it was right to prognosticate lung cancer incidence from time-series and improve forecast precision through developing combination models. The result of combination seemed close to ARIMA models which suggested that it could serve as a chief way to forecast the incidence of lung cancer.