Cassagne E, Caillaud D, Besancenot J P, Thibaudon M
Climat et Santé, Centre d'Epidémiologie de Population, Faculté de Médecine, Dijon, France.
Eur Ann Allergy Clin Immunol. 2008 May;40(1):14-21.
Pollens of Poaceae are among the most allergenic pollen in Europe with pollen of birch. It is therefore useful to elaborate models to help pollen allergy sufferers. The objective of this study was to construct forecast models that could predict the first day characterized by a certain level of allergic risk called here the Starting Date of the Allergic Risk (SDAR). Models result from four forecast methods (three summing and one multiple regression analysis) used in the literature. They were applied on Nancy and Strasbourg from 1988 to 2005 and were tested on 2006. Mean Absolute Error and Actual forecast ability test are the parameters used to choose best models, assess and compare their accuracy. It was found, on the whole, that all the models presented a good forecast accuracy which was equivalent. They were all reliable and were used in order to forecast the SDAR in 2006 with contrasting results in forecasting precision.
禾本科花粉是欧洲最具致敏性的花粉之一，与桦树花粉相当。因此，构建模型以帮助花粉过敏患者是很有用的。本研究的目的是构建预测模型，该模型可以预测以特定过敏风险水平为特征的第一天，在此称为过敏风险起始日期（SDAR）。模型源自文献中使用的四种预测方法（三种求和法和一种多元回归分析）。这些方法于1988年至2005年在南锡和斯特拉斯堡应用，并在2006年进行测试。平均绝对误差和实际预测能力测试是用于选择最佳模型、评估和比较其准确性的参数。总体而言，发现所有模型都具有良好且相当的预测准确性。它们都很可靠，并用于预测2006年的SDAR，但预测精度结果有所不同。
