RsuMatcher: A personalized rsum-job matching system Academic Article uri icon

abstract

  • 2016 Elsevier Ltd. All rights reserved. Online jobs search through popular websites are quite beneficial having served for many years as a prominent tool for job seekers and employers alike. In spite of their valuable utility in linking employers with potential employees, the search process and technology utilized by job search websites have not kept pace with the rapid changes in computing capability and machine intelligence. The Information retrieval techniques utilized by these websites rely primarily on variants of manually entered search queries with some advanced similarity metrics for ranking search results. Advancements in machine intelligence techniques have enabled programmatic extraction of pertinent information about the job seeker and job postings without active user input. To this end, we developed a resume matching system, RsuMatcher, which intelligently extracts the qualifications and experience of a job seeker directly from his/her rsum, and relevant information about the qualifications and experience requirements of job postings. Using a novel statistical similarity index, RsuMatcher returns results that are more relevant to the job seekers experience, academic, and technical qualifications, with minimal active user input. Our method provides up to a 34% improvement over existing information retrieval methods in the quality of search results. In addition however, RsuMatcher requires minimal active user input to search for jobs, compared to traditional manual search-based methods prevalent today. These improvements, we hypothesize, will lead to more relevant job search results and a better overall job search experience for job seekers. As an alternative to the fragmented organization-centric job application process, job recruitment websites offered the promise of simplifying and streamlining the job search process. However, these websites offer limited functionality using generic and simplistic information retrieval methods, which being non-domain lead to a poor and frustrating search experience. In this paper, we present RsuMatcher, a personalized job-rsum matching system, which offers a novel statistical similarity index for ranking relevance between candidate rsums and a database of available jobs. In our experiments we show that our method offers a 37.44% improvement over existing information retrieval methods in the quality of matches returned.

published proceedings

  • Expert Systems with Applications

altmetric score

  • 0.5

author list (cited authors)

  • Guo, S., Alamudun, F., & Hammond, T.

citation count

  • 41

complete list of authors

  • Guo, Shiqiang||Alamudun, Folami||Hammond, Tracy

publication date

  • October 2016