Interoperable Subject Vocabulary Service

From Xcri

[edit] Introduction

Since the inception of the XCRI strand of work there has been an increasing expectation that XCRI will address vocabularies, particularly subject vocabularies, even though this is not directly within its remit. A common method of searching across repositories of learning opportunities has been the ‘holy grail’ of course information systems, so that learners do not have to grapple with different search mechanics in otherwise very similar systems. The project will take a sample subject area from several subject classification systems and lists of keywords and map them to a common spine vocabulary, probably the Learndirect Classification System. It will develop a web-based demonstrator of the full service, and a working prototype for the sample subject area. The prototype will show how a user can select a search term from a keyword list or subject classification system and obtain search results from a different keyword list or classification system used in the course information system. The project will investigate and compare the quality of results using the ‘native’ system and the enhanced subject search service.

Final Report

Interoperable subject vocabularies for better search results

[edit] Advanced Search Algorithm Protocol (ASAP) Prototype

The ASAP prototype is published here. For login and password, or for a demonstration, please contact Alan Paull at alan@alanpaull.co.uk.

The approach of the ASAP search is to map XCRI-CAP format courses onto the system's bespoke thesaurus using many classification systems and keyword lists. By using a number of different classification methods ASAP provides both a common, generic approach to mapping courses onto the thesaurus and a tailored approach for institution and even course specific mappings. As each course file is imported into the ASAP system, it is matched against the relevant thesaurus terms using one or more of the classification systems, and the thesaurus term is added to the course record.

When a user searches using ASAP, they begin by entering a search term. That search term is then compared to the thesaurus and used as an 'entry point' into the structure of the thesaurus. The system then follows each relationship through the thesaurus to build a specific 'sub-thesaurus' for that search term. The sub-thesaurus is then used to query the courses held in the database, and all courses linked to the nodes within the sub-thesaurus are retrieved. Each course is tagged with its relationship to the original search term node, and scored depending on its relevance and distance from the original node. The resulting data set is then returned for presentation to the user.

We have provided a script that users can follow to see the ASAP mechanism and to compare it with conventional subject searching.