Grant: Engage Grants Program

Search is a vital part of many information-centric activities. Traditional approaches for supporting search tasks focus on keyword-based matching and ranked listed representations. This works well when the searcher can accurately describe what it is that they are seeking, and when the terminology used in the documents is the same as that used by the searcher. However, such keyword-based approaches become less effective when there is a large degree of ambiguity within the searcher's query, the collection being searched, or both. For example, searching within a collection of job applications for a particular skill will be successful if both the recruiter and applicant use the same terms to describe this skill, and will fail if there is a mismatch in their terminology. As a result, it is necessary for searchers to perform a series of search tasks based on their knowledge of the domain, and expend significant effort in synthesizing the ranked lists of search results.The goal of this research is to develop and evaluate a novel approach for addressing the term mismatch problem within the context of job application search tasks. HireGround has developed software to collect a large amount of data on job titles and links to the job postings. We will leverage this information in the construction of a custom ontology that will hold the relationships between the skills listed in the postings and the job titles associated with these skills. This ontology will then be used to support the resume search process through automatic query expansion, and a visual organization of the search results based on the extracted skills. The benefit of such an approach is that the searcher can initiate the search process with a few simple terms that describe the skills they are seeking, and allow the ontology-driven search engine to match this to the job applications at a conceptual level. The search results will be visually organized using the skills and associated job titles extracted from the ontology, enabling interactive filtering and exploration among the broad collection of search results to re-focus the results on those that match the skills that are being sought.