A Hybrid Machine Learning Approach for Fake Job Posting Detection: Integrating Naive Bayes and Logistic Regression Models

The recruitment sector's digital transformation greatly eased the process of job hunting, and the use of online job portals increased, but it also paved the way for malicious players to put up scam job listings on these websites. These scam listings can result in job scams, data breaches, and even emotional trauma to the prospective employees. In order to address the growing menace of job scam postings, this study presents a hybrid system to detect spam job postings in a more precise and dependable manner. The system couples the Naive Bayes-based probabilistic capabilities, especially useful when applied to text- intensive data, with Logistic Regression's predictive ability, which excels in the case of binary class problems. By using the two techniques combined, the system effectively encompasses linguistic traits typical of spam job postings, along with better performance in the task of classification. This study uses a high-quality, real-world dataset composed of heterogeneous job-related features, including unstructured text fields like job titles and company info, and structured attributes like employment type, locality, and business type. Preprocessing techniques are added to augment the analysis of the unstructured text, including tokenizing, stopword removal, stemming, and vectorization using TF-IDF. Both feature scaling and feature set gathering are also used to reduce the dimensionality of the data and increase the efficiency of the model. Multiple metrics of classification are assessed using the combined system, including precision, recall, F1-score, and ROC-AUC. Experiments exhibit the comparative superiority of the combined system, especially in precision and recall in detection of even slight indicators of spam postings. This study provides a viable and scalable measure to raise the security of the recruitment system online by facilitating preliminary detection of scam job opportunities and helping in augmenting security, trust, and, more importantly, the safety of millions of job seekers it is likely to integrate alongside existing job portal systems, thus enhancing safety and trust amongst its numerous users.