Vocabulary Mapping
An application that would analyze a keyword search and generate some form of tag cloud with recommended controlled vocabulary terms. If this were somehow able to visually convey the number of records under each term, and the interrelation between them, I think that would be beneficial. I’m trying to describe something that would be a cross between a traditional tag cloud, like that of Delicious, and something possibly like either LivePlasma.com recommendation engine or AquaBrowser’s sidebar.
Value for Users
This type of application would push users into using controlled vocabulary that they so rarely seem to know exists. This means of browsing could help not only to indicate ideal search terms for the initial idea, but can also create a form of serendipitous discovery where related ideas are presented to the searcher by showing the relationships between terms.
Application Users
I envision this app having two typical users. The first would be undergraduate students using this app to improve the quality of their research and to retrieve exhaustive results on a topic. I could envision graduate students using this tool to narrow or broaden areas of study, and to find related areas of study for dissertations. I think that the sciences are a field that most needs this, as there are so many different ontologies and thesaurii available.
Loading ...
6 Comments
Excellent idea to discover ‘hidden’ information (a similar approach, but limited to pubmed, is taken by gopubmed which takes a search in pubmed and maps it back to mesh). If I understand this correctly, the suggested idea would extend post search filtering capabilities of Scopus and take that to a new level.
I see this as a winner at all levels. Keyword searching, in and of itself is not very useful. It tends, in my opinion, to often lead users astray and away from the really useful gems that are available. By making the controlled vocabulary user functional all would benefit.
Anything that helps endusers discover and use controlled vocabulary is a good thing. It would really expand their understanding of searching the literature.
I think this is another application which uses the display of information much like to concept maps. So far it is possible to regroup keywords from terms chosen categories and disciplines of the title, abstract and citation of work and visualize the number of records that are isolated or groups establishing relationships between them can arise where new information. Of course that from the searches may be creating a kind of controlled vocabulary based on frequency of use of the various terms and any recommendations for users of the application at least so I understood the idea and I’m completely agree with it
A good idea, and one that has had some traction in library search, federated search and discovery services over a few years. There are a few search services which will analyze results and produce a tag cloud of terms found in the articles. At least one can represent these as a tag cloud with font size representing the tag/term weight and the number of hits next to it. These terms can then be selected to build new searches, rinse, and repeat. Problems arise from the lack of data (most searches return metadata, which typically does not include an abstract, so the analysis is left working on the title – which is poor and often quirky), and the actual normalization and then analysis processing which much be done. As an add-on to an existing search interface these processing hits could likely be avoided, but then the app is tied to that system…
Identifying what the controlled vocabulary terms are for a given database is usually harder than it should be and this idea could make identifying good keywords a lot easier. It would also be really nice if there was a way for a user to circle a cluster of terms to refine a search instead of needing to sort through a list of check boxes.