多元聚合分类方案

取得对样品（或逆分析中的要素）的客观分类（型），对于提高天气预报的效果有着十分重要的意义，遗憾的是目前尚未见到能够获得合理分类结果的可行客观分类方法。本方案是以聚合为策略，以计算机为工具，通过逐次全面比较所给样本资料各种情况下的分类，试图在这方面进行的一项尝试，方案目前已取得下列成果:① 提供出一种新的分类思路和方法；② 分类客观且有较好的效果；③ 对小样本可确保取得合理分类。

Multiple Aggregation Classification Method

Classifying samples into types is very important to the correctness of weather forecasts. It is the pity that there is probably no classificational methods which can be used to obtain a satisfactory result. The present paper attempts to have a try in this field by aggregating, computing and comparing various classifications with the given data. The results show that (i) it is a new type of classificational methods and (ii) it is objective and can be used to obtain fairly good results especially with small sample size.