A Comparative Study on Fake Job Post Prediction Using ML Technique

The growth of online recruitment platforms has created many job opportunities, but it has also increased the number of fake job postings. Fraudulent advertisements often try to collect personal information or money from job seekers by offering false promises. To solve this problem, this paper presents FakeJobPrediction, a hybrid system for detecting fake job posts. The system combines a Deep Neural Network (DNN) model with a rule-based engine to classify job listings as real or fake. The model is trained using the EMSCAD dataset from Kaggle along with additional synthetic data to improve performance. Natural Language Processing (NLP) techniques such as text cleaning, tokenization, stop-word removal, and TF-IDF vectorization are applied to process job descriptions. The DNN identifies complex text patterns, while the rule-based engine detects suspicious indicators like unrealistic salaries, vague content, and generic contact details. The hybrid approach improves detection reliability and achieves an accuracy of 98.67% on the test dataset. The system is deployed using a web-based architecture that provides real-time predictions with confidence scores. FakeJobPrediction offers a practical and effective solution to protect job seekers from online recruitment fraud.