Abstract

With the advent of the internet and globalization, most services have become automated, including the recruitment process. E-recruitment systems have gained popularity due to their ability to automate various steps of recruitment. However, while some systems filter and screen CVs for skills, they do not assess a candidate's personality, which is crucial in determining their compatibility with a company's culture and practices. This compatibility leads to effective recruitment and happier, more productive employees who stay at the company longer. Due to the lack of a personality detection mechanism, recruiters spend a significant amount of time conducting multiple interviews to assess candidate fit. To address this issue, the authors propose a system that uses fine-tuned BERT to detect a candidate's MBTI personality from their CV or personal statement. The proposed system achieved an average accuracy of 72.14% per class and an AUC range of 0.85-0.90, following pre-processing of the Kaggle MBTI personality types dataset to address class imbalance.