Study on Automated Resume Information Extraction and Candidate Profiling

Recruiters and placement cells get hundreds of resumes that come in all sorts of formats and layouts, making it tedious to read through them, inconsistent and difficult to repeat. Automated resume parsers remove this problem, but most of them stop at extracting a handful of simple details like a name, an email address or a list of keywords, leaving the recruiter with a flat set of fields instead of a useful picture of the candidate.This study examines automated resume information extraction with an explicit focus on candidate profiling. It looks at how the name, email address, mobile number, skills, educational qualifications, work experience and projects can be extracted from uploaded PDF and DOCX resumes and then combined into one structured candidate profile. A hybrid framework is proposed. The text is extracted using the Apache PDFBox and Apache POI tools, normalized, and split into sections with a header dictionary and layout hints. The patterns, curated dictionaries, and a statistical named entity recognition model from Apache OpenNLP library are applied to fill each section’s fields. The profile builder links each skill to the experience and project fields, calculates the total of experience by combining the ranges, and outputs a completeness score with a confidence level for each. The paper describes the design of each stage, the data model, the Java-based implementation plan with the Spring Boot and Hibernate frameworks, and an evaluation procedure with precision, recall, and F1 measures based on annotated resumes. The main point is that the creation of the profile view with interconnected fields of skills, projects, and experience allows to transform the text of the resume into screening and search data.