Classification: Purposes, Principles, Progress, Prospects
Classification: Purposes, Principles, Progress, Prospects
复制标题
DOI：
10.1126/science.185.4157.1115
复制
发表时间：
1974-09
期刊：
影响因子：
56.9
通讯作者：
R. Sokal
中科院分区：
文献类型：
--
作者：
R. Sokal
分类的原理和程序之间存在密切的相互关系，该领域的现代工作受到电子计算机发展的深刻影响。除了描述自然系统以及实现记忆的经济性和易于操作之外，分类的主要目的是描述相似对象组的结构和关系。尽管许多分类工作已经应用了实际目标，但成功的分类会产生科学假设。对多类群的接受是一个重大的概念进步，并直接导致基于许多同等权重特征的分类。计算机分类数据的规范将增强客观性，但不能消除文化和主观偏见。分类技术包括聚类分析和排序，并且最近已经阐述了多种表示分类的方法。通过将图论应用于一些分类问题，重建进化分支序列已经成为可能。计算机分类已成功应用于广泛的学科。
There is an intimate interrelation between principles and procedures in classification, and modern work in this field has been profoundly affected by the development of electronic computers. Besides the delineation of natural systems and the achievement of economy of memory and ease of manipulation, the primary purpose of classification is the description of the structure and relationship of groups of similar objects. Successful classifications generate scientific hypotheses, although much classificatory work has applied, practical goals. The acceptance of polythetic taxa is a major conceptual advance and has directly led to classifications based on many, equally weighted characteristics. The specification of data for classification by computer will enhance objectivity but not eliminate cultural and subjective biases. Techniques of classification include cluster analysis and ordination, and numerous ways of representing classifications have been elaborated recently. By the application of graph theory to some classificatory problems it has been possible to reconstruct evolutionary branching sequences. Computer classification has been successfully applied across a broad range of disciplines.