Resume Parsing Framework for E-recruitment

Modern approaches to improve networking and communication have given ways to the advancement of recruitment process through the development of e-recruitment recommender systems. The increasing expansion of internet- based recruiting has resulted in a large number of resumes being stored in recruitment systems. Most resumes are prepared in a variety of styles to attract the attention of recruiters, including different font sizes, font colors, and table formats. However, data mining operations such as resume information extraction, automatic profile matching, and applicant ranking are immensely affected by the variety of formats. Rule-based methods, supervised methods and semantics-based methods have been introduced to extract facts from resume accurately, however, these methods heavily depend on large amounts of annotated data that is usually difficult to collect Furthermore, these techniques are time-intensive and bear knowledge incompleteness that strongly affect the accuracy of resume parser. In this paper, we present a resume parsing framework that handles the limitations faced in the previous techniques. At first, the raw text is extracted from resumes and blocks are separated using text block classification. Furthermore, the entities are extracted using named entity recognition and enriched using ontology. The proposed resume parser accurately extracts information from resumes that directly contributes towards the selection of best candidate.