Prediction Of Fake Job Ad Using Nlp-Based Multilayer Perceptron

Abstract

Today, technological and industrial developments have created new and diversified career opportunities for job searchers. Based on suitability, experience, qualification, time, etc., job seekers find out their options with the help of the advertisements for these job offers. The recruitment process is now influenced by the power of social media and the internet recruitment process. To share the job details in electric media Social media and advertisements created new and newer opportunity. To share the job ads, the growth of opportunity has increased the number of fake postings that may cause harassment of job seekers. Due to the security and consistency of their professional, academic, and personal information, people may show interest in new job postings. To attain people’s reliability and belief, electronic and social media face an extremely hard challenge—the true motive of valid job postings. To make life easier and more developed, there are technologies around us but for professional life, they create an unsecured environment. For recruiting new employees, employees will be a great advancement if the job ads can be properly filtered, predicting fake job ads. This paper proposes various data mining techniques and classification algorithm such as random forest classifiers, NB classifiers, SVM, DT, K-nearest neighbors, and multi-layer perceptron, for the prediction of whether job advertisements are real or fraudulent. On the Employment Scam Aegean Dataset (EMSCAD) we have examined 18000 samples. For this classification task as a classifier, the deep neural network performs great