Towards automatic hypertextual representation of linear texts (Q4376799)

It is a fact that traditional information retrieval (IR) techniques alone cannot cope with the steadily increasing volumes of textual information. This paper presents a survey of existing and implemented, as well as of projected approaches towards automatic linking, which constitutes the basis for sensible addition to conventional IR methods: associative searching by means of hypertext links. Describing the mechanisms for the automatic transformation of linear texts into a nonlinear form, the paper shows and compares the strengths and weaknesses of different approaches to computing hypertext links: extraction of nonlinear structures, statistical evaluations, natural-language parsing, high-level modeling, monitoring user actions, link inheritance and accumulation. For the entire collection see [Zbl 0879.00058].