IMPROVING STUDENT EMPLOYABILITY BY UTILISING SEMANTIC ANALYSIS OF COURSE DATA

This paper describes the development of a toolkit based on the semantic analysis of course data using the eXchanging Course Related Information – Course Advertising Profile (XCRI-CAP) information model. Based on the XCRI-CAP 1.2 standard, our research focused on providing the means for semantically comparing course descriptions and curriculum documentation. The paper provides an overview of how the JISC funded MUSAPI (Musket-Salami Project Integration) project resulted in the creation of a web service allowing
(i) raising awareness regarding the architecture of job profiles and the employability terminology used in career services,
(ii) identifying potential job opportunities for student graduates and/or university applicants,
(iii) ranking ‘real’ job opportunities filtered according to a range of criteria and
(iv) mapping job opportunities to courses in higher education institutions offering the necessary skills and knowledge.

The paper initially describes how the MUSAPI project integrated web services developed as part of the SALAMI project in Nottingham University with the MUSKET algorithms generated from Middlesex University. The scope of the project was to provide further support at national career services by providing more sophisticated means for browsing jobs. Initially, the use of a thesaurus enables users to increase their awareness of how generic job titles can be inclusive of various job profiles leading to truly different careers. The maintenance of such a thesaurus allowed the definition several job profiles and the creation of the architecture of relationships between jobs in specific sectors. With the use of a spider algorithm, ‘real-life’ job opportunities are retrieved from a number of job-seeking websites, including the national career services. The search criteria are based on user-provided keyword sets, enriched with the use of WordNet. The retrieved job opportunities (each one corresponding to a unique, available job) are ranked against a number of criteria, ranging from location and salaries to skills and experience required. A number of visualisation options are available for classifying job opportunities and performing the necessary filtering and ranking.

The MUSKET (Middlesex University Skills and Educational planning Tools) algorithms are used to compare job opportunities against the user selected keyword sets. Finally, the same algorithms are used to retrieve suitable courses from a centralised repository, matching the job opportunity descriptions. The scope of the MUSKET project was to calculate the semantic similarity between the various courses available, by comparing similar aspects of the course documentation (e.g. learning outcomes, topics covered, course aims, pre-requisites). Hence, the user can assess which university courses available are relevant to a job opportunity and decide whether a specific course is sufficiently similar to provide the necessary skills for such a career.