用于界面设计的文献识别模型

Document Discrimination Modeling for Interface Design

文献分析的自动化方法在因特网环境中尤其必要。采用若干术语进行标引确定了一个类目中的组成成分 ,但抑制了对这些成分的深入理解和辨析。文献集合中一个术语的平均频率是一个单纯的功能 ,并且是一个有趣的、可视化的和友好特性的功能。一个为用户开发的界面 ,使得用户容易的操纵和发现内容的上下文 ,内容定位了术语在全文中的位置。本文描述了一个实验 ,它确定了这个方法的一些有用的可能性。

Automatic methods for document analysis are particularly necessary in the Internet space. Indexing terms identify members of a class but constrain discrimination amongst them. The average frequency of a term over the document set is a simple function with an interesting visual friendly property which an interface could exploit for users to easily manipulate and find the context of content bearing terms in the full text. This paper describes an experiment that confirms some useful possibilities of this approach. Some implications for visualization are listed to help users analyze document sets and in the construction of contextual information.