Explainable Fake Job Posting Detection Using OCR, Machine Learning, and AI Reasoning

Abstract: The increased availability of online recruitment sites has increased the possibility of fake job ads targeting potential employees. This paper proposes a transparent solution for the detection of fake job ads that utilizes Optical Character Recognition, a LRND ensemble classifier, SHAP, and an AI reasoning module. The proposed solution utilizes the EasyOCR library for Optical Character Recognition, TF-IDF for feature engineering, and a LRND ensemble classifier for the detection of fake job ads with an accuracy of 93.5%. SHAP is used for feature interpretation, and an AI reasoning module is used for providing explanations for the detection of fake job ads. The proposed solution is developed as a web application using the Flask framework.

Keywords: Fake job detection, Optical Character Recognition, TF-IDF, LRND classifier, SHAP, Flask, machine learning, fraud detection