【摘要】 针对当前有关疾病“三间”分布的聚集性分析这类方法中存在着的通病，本文尝试运用趋势面分析方法来对分布建模．为此，我们廓清了广于“聚集性”的“趋势性”之概念，将二元多项式的趋势面模型拓广到1～3元；依据设立“对照”原则，导出了与人口的人群、空间、时间分布及其联合相关分布相对应的7个疾病的分布，并且利用其相互间的可对比关系可以作出有关地理相关性的推断．同时，对各分布的趋势面模型拟合改用逐步回归代替完全回归．用这套建模方法成功地对一个实例作了研究．最后，讨论了建模中的有关问题，特别指出了设立“对照”原则在数学应用的方法论上的普遍意义，即数学应用的思维模式与实验科学的思维模式的一致性．更多还原

【Abstract】 To counter common failing existing in this kind of clustering analysis methods about distributions of disease in crowd-space-time at present,this text tried applying trend surface analysis method to modeling the distributions.For this purpose,we cleaned up the concept of "tendency"which is more general than "clustering",and extended trend surface model for bivariate polynomial to 1～3 variate ones.And acrrording to the principle of establishing"contrast",we infered 7 distributions of disease corresponding to distributions of population in crowd-space-time and its joint-correlative distributions,and could deduce the geographical dependency by means of the comparable relations amongst them. Simultaneously,we substituted stepwise regression analysis for complete regression analysis in fitting to trend surface models of these distributions.This modeling method successfully was applied to a living example study.Lastly,we discassed some problems about modeling,and especially pointed out univ rsal significance for the principle of establishing "contrast"in the methodology for applying mathematics,i.e.the identity of thinking pattern of applying mathematics and thinking pattern of experimental science.更多还原