Application of Iterative Machine Learning in Predicting Fake Documents in Job Applications

Abstract

The increase in the rate of industry liquidation and unproductiveness resulting from false persons’ recruitment through fake document assessment during the job application is a great concern. Most companies have gone bankrupt whereas some have shut down untimely due to unqualified staff recruited with fake documents. This paper proposed a machine learning (ML) model that predicts fake documents in job applications. The machine learning model was designed with iterative learning algorithms based on a comparative analysis of applicants’ data and online information from the alma mater discussed in this article. An iterative similarity convergence threshold of 0.5 was set to ensure the prediction accuracy of the model. Compared documents that are less than the defined threshold are considered fake and not original. The iteration results of several trained datasets presented in Tables 4 and 5, filtered fake documents. These datasets were trained, tested, and validated using the MATLAB application. Significant results achieved, revealed the efficient performance of the model in filtering fake documents in job applications.