Does cross validation provide additional information in the evaluation of regression models?
Does cross validation provide additional information in the evaluation of regression models?
复制标题
DOI：
10.1139/x03-022
复制
发表时间：
2003-06-01
期刊：
影响因子：
--
通讯作者：
Kozak, R
中科院分区：
文献类型：
--
作者：
Kozak, A;Kozak, R
利用7个数据集、2个立木体积估算模型和1个树高-直径模型进行的详细研究表明，通过交叉验证或双交叉验证的模拟，可以很好地估计从回归模型直接计算的拟合统计量和不拟合统计量。这些结果表明，通过数据分割和双重交叉验证的交叉验证在评估回归模型的过程中提供了很少的额外信息。
A detailed study using seven data sets, two standing tree volume estimating models, and a height-diameter model showed that fit statistics and lack of fit statistics calculated directly from a regression model can be well estimated using simulations of cross validation or double cross validation. These results suggest that cross validation by data splitting and double cross validation provide little, if any, additional information in the process of evaluating regression models.