组合模型在肺癌死亡率预测中的构建和应用
Establishing a Combination Model in Predicting Mortality of Lung Cancer
-
摘要: 目的 用3种预测方法对肺癌死亡率进行拟合,选择最优方法,建立组合模型并进行预测。方法 用动态数列、joinpoint回归、指数平滑对2001~2013年金昌队列肺癌死亡率进行拟合,选择单项模型形成组合模型进行拟合并预测。组合模型权重系数的计算基于算术平均法、方差倒数法、均方误差倒数法、简单加权平均法。结果 单项预测模型以指数平滑法拟合精度最高,为79.67%,Joinpoint线性回归拟合精度为74.27%。指数平滑与Joinpoint线性回归进行组合,其中以算术平均法及均方误差倒数法预测效果最好,拟合精度分别为86.87%、85.80%。结论 组合模型优于单项预测法,可以用于肺癌死亡率预测。Abstract: Objective To identify a good combination model for predicting the mortality of lung cancer. Methods Mortality data of lung cancer from 2001-2013 were used to test three prediction model: dynamic series, exponential smoothing, and Joinpoint regression. Weight coefficients of the combination models were calculated using the arithmetic average method, the variance inverse method, the mean square error inverse method, and the simple weighted average method. Results The exponential smoothing model had the highest accuracy (79.67%) of prediction, followed by the Joinpoint linear model (74.27%). The combination of these two models resulted in better results. The arithmetic average method and the mean square error inverse method had the best prediction, with an accuracy of 86.87% and 85.80%, respectively. Conclusion The combined model has higher accuracy than the single models in predicting the mortality of lung cancer.
© 2018 《koko体育app
学报(医学版)》编辑部 版权所有
开放获取𝓡 本文遵循知识共享署名—非商业性使用4.0国际许可协议(CC BY-NC 4.0),允许第三方对本刊发表的论文自由共享(即在任何媒介以任何形式复制、发行原文)、演绎(即修改、转换或以原文为基础进行创作),必须给出适当的署名,提供指向本文许可协议的链接,同时标明是否对原文作了修改;不得将本文用于商业目的。CC BY-NC 4.0许可协议详情请访问