Web App For Predicting Fake Job Posts Using Ensemble Classifiers

Taking advantage of the internet, cheating people online has become common these days, there are numerous types of scams done online, employment scam is one of them, where fake job postings are published online to attract people who are unemployed and looking for work. By becoming victims of these fake job postings, many people lose their personal information and money. This needs to be dealt with intelligence and caution. In this paper, we have proposed a comparative study using different ensemble classifiers which work effectively on this problem and help to predict whether a particular job is genuine or fake. We have used five different ensemble classifiers: Random Forest, LightGBM, XGboost, Catboost, and Extra tree. EMSCAD(Employment Scam Aegean Dataset) which contains 17880 job posts is used to train the models. We have created a web application to make it further easy for users to take advantage of this facility. XGBoost classifier has achieved 96.8% accuracy. which is highest among other classifiers also it was observed that the Light GBM classifier is faster than the other classifiers.