Online Recruitment Fraud (ORF) Detection Using Deep Learning Approaches

ABSTRACT:

Online recruitment platforms have revolutionized job searching, but they also present significant challenges due to the rise in fraudulent job postings that deceive job seekers and lead to financial and personal harm. This study proposes an advanced deep learning-based approach to detecting online recruitment fraud (ORF). The research utilizes the publicly available Fake Job Posting Prediction dataset, which contains real and fraudulent job postings, for training and evaluation. Existing models such as BERT and RoBERTa have demonstrated strong performance in text classification tasks. However, to further enhance accuracy and robustness, we introduce a hybrid approach leveraging GPT-2, XLNet, and LSTM. GPT-2 and XLNet, with their transformer-based architectures, are employed to capture contextual and semantic relationships in job descriptions, while LSTM is used for sequential pattern learning to improve fraud detection. The system is designed to effectively distinguish between real and fraudulent job postings, minimizing false positives and false negatives. The performance of the proposed models is evaluated using standard classification metrics such as accuracy, precision, recall, and F1-score. Experimental results indicate that our approach significantly improves ORF detection, offering a reliable solution to safeguard job seekers from fraudulent employment opportunities. This research contributes to enhancing trust and security in online recruitment platforms.

Keywords: Online Recruitment Fraud Detection, Deep Learning, GPT-2, XLNet, LSTM