Fake Job Recruitment Detection using Data Mining Techniques

On-line recruitment sites have been defined as altering the searching behaviour of the job seekers on their job opportunity availability, but the challenges are serious and include but are not limited to, the rise of fake job advertisements that are defrauding the job seekers and killing their organisation reputation, money, and lives. The provided paper presents an algorithm to identify online recruitment fraud (ORF) which is more effective than the previously implemented algorithms based on the banking industry data. Training and evaluating the research are performed with Fake Job Posting Prediction that is publicly available and uses real and fake postings. The existing models such as BERT and RoBERTa have performed well in text classification tasks. GPT-2 and XLNet architectures based on transformers are used to learn contextual and semantic associations in job descriptions but LSTM is used to learn sequences to improve fraud detection. The system will be resistant to fraud in context of real job posting, and fake job posting with a predetermined limit in terms of false positive and false negative. Results of the experiment: It is discovered that the model performances proposed are approximated on the basis of the conventional measures of classification, such as accuracy, precision, recall, and F1-score, and are much more effective as to insure job seekers against fraudulent jobs. The research would contribute to build trust and security on the internet as regards to recruiting.