Determining a Proper Text Similarity Approach for Resume Parsing Process in a Digitized HR Software

Resume parsing is one of the costly phases of a recruitment process. This phase has been alleviated in digitized human resources recently by using text processing approaches between a job advertisement content and resume of applicants. For this purpose, performing a text similarity calculation is one of the most commonly used approaches. However, there are lots of similarity calculation models and most of them are not targeted a recruitment process. Moreover, a subjective assessment of such approaches is required to provide a proper text processing in such a specific problem domain. Thus, in this paper, we offer to evaluate different similarity score calculation approaches through a recruitment case study with the help of a statistical assessment. For this purpose, a computer-aided resume evaluator on a set of resumes is proposed, a human evaluation on the same set of resumes is performed by the professions and the correlation between the outcomes is sought out. As a conclusion, a discussion among different similarity score calculation approaches available for resume processing is presented to find out a proper computer-aided resume evaluator for digitized human resources.