Text Mining and Scholarly Publishing

Use the link to access the full text report opening as follows: "What is text mining? How does it relate to data mining? Why do people want to do text mining? How does it work? What do publishers need to do to support text mining? There are many questions swirling around the topic of text mining of the scholarly literature. At the time of writing, text mining is generating a frenzy of debate in the scholarly publishing world. There is the usual misunderstanding, over-enthusiasm and unrealistic expectations that are associated with technology hype.  There is no universally agreed definition of text mining. This is partly because it is being used by different communities for different purposes. Each community has its own preferred scope and definition of text mining. This can lead to disagreements over where information extraction finishes and text mining starts 

or the difference between text mining and data mining. This paper aims to disentangle the topic and to clarify the underlying issues for the general, non-expert reader. The scope of this paper is the text mining of scholarly journals and books. The word 'text' is used to describe the content of these sources and 'publishers' refers to scholarly publishers only. The focus is on what scholarly publishers can do to make their content more machine-accessible, although it is hoped that others will find this paper helpful, for instance for those who are responsible for making policy in this area, whether it be at trade or government level, researchers not currently engaged in the area, librarians, and other interested parties ..."