A Hybrid NLP-Expert System Framework for Heuristic Scoring and Fairness-Driven Resume Evaluations

Automated resume screening tools have become popular because they save time and make the hiring process more efficient. However, these tools have sparked concerns around fairness and transparency, as there is a risk that they could reinforce existing biases. To address this issue, we are proposing a hybrid framework that combines straightforward content evaluation with sentiment analysis and explicit fairness rules. Our approach uniquely integrates a well-known sentiment analysis tool called VADER, which helps gauge the positivity of language used in resumes, alongside a simple measure of information each resume provides. Additionally, the framework incorporates an expert system that implements specific fairness rules to identify and adjust for potential biases. In the experiment with a real set of resumes, the framework confirmed its ability to effectively rank candidates based on their qualifications while also identifying and reducing bias in the evaluation process. The results revealed that most resumes received high fairness scores, showing minimal bias issues. Overall, our hybrid framework demonstrates a practical way to make automated resume evaluations more transparent, fair, and balanced by combining NLP tools with human-inspired fairness rules.