Methodology for resume parsing and job domain prediction

Abstract

With the rapid growth in on-line based recruiting system, candidates apply for job on web portal by uploading their resumes. Due to internet based recruiting systems, candidates participate in large volumes and hence it becomes a challenge task for recruiter to filter candidates for the required role. The resumes uploaded by the candidate are varied in format such as font, colour, font size, etc. and it is difficult for the recruiters to find the best match for a job role. Natural Language Processing (NLP) helps to deal with such problems and help recruiters to extract detailed information of the candidates required to carry forward their candidature. In this work, we propose to use named entity recognition of Stanford CoreNLP system to extract information relevant for recruiting process. Moreover, on the basis of skills set of candidate, the resume of the candidate is assigned a genre such as Computer Science, Statistics, Business Development etc. In this paper we propose to design an intelligent resume parser system capable of converting the unstructured data into a structured format which enables the recruiter to filter the right candidates for the desired job role. The overall resume prediction of our system is 91.47%.