temulenbd/text-representation-comparison-job-recommender

Project Author: Temuulen Bulgan

Topic: A Comparative Evaluation of Text Representation Techniques for Content-Based Job Recommendation System

Summary: To address information overload in human resources, a study was conducted on Job Recommendation Systems (JRS), involving a review of nearly 300 research papers. A gap was identified in evaluating text representation techniques within JRSs. A custom JRS was developed to compare three methods—TF-IDF, Word2Vec, and BERT—through both offline and online experiments. The research introduced a new framework for assessing the quality of content-based JRSs, offering a foundation for future benchmarking and development.