Detection of Fake Online Recruitment Using Machine Learning Techniques
Tejasva Bhatia, Jasraj Meena · 2022
There are a numerous amount of job postings on the internet and sometime these vacancy postings turns out to be fake. Even on the reputed and trusted job advertising platforms, people fall prey to these fake advertisements . After selection in the job, hiring people start demanding for money and details of bank account. Good number of candidates gets duped and lose loads of money and sometimes even their current jobs. So it would be very helpful to identify if the job listed on the website is real or fake. In this paper, we have used machine learning to detect fraud job vacancy postings.In our proposed ML technique, we have applied two data balancing techniques namely "Adaptive Sympathetic" and "Synthetic Minority Oversampling Technique" in combination with "Term Frequency-Inverse Document Frequency" which if a feature extraction method. However, in the literature, some research have used "Bag of Words" for extraction of features which could just count the no. of times of the word appeared whereas the technique used in this research work (i.e TF-IDF) also provides the importance of words. The public "Employment Scam Aegean Dataset" (EMSCAD) was used which contained around 18000 job listings out of which about 800 are fraud listings. We have used two machine-learning models such as Random Forest and k- nearest classifiers to detect the online job is face or real. We have compared the performance of proposed models with existing models and it performs better in terms of recall, precision and f1-scores.