Application of natural language processing for fake job posting detection

Gadikota, Vaishnavi (2025) Application of natural language processing for fake job posting detection. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (658kB) | Preview
Preview	PDF (Configuration Manual) Download (590kB) | Preview

Abstract

The growing trend of counterfeit job ads on online job market sites has posed serious threats to job seekers and has drastically caused a loss of faith in online job markets. These fraudulent postings have not always had identifiable organisations, have been in imprecise language and have sought to gather personal data or deceive applicants. This project has managed this emerging threat by coming up with a Natural Language Processing (NLP)-based fake job posting detection system that can determine deceptive advertisement through linguistic and metadata-based patterns. The study aimed at evaluating NLP and machine learning models to detect fraudulent job offers, determining linguistic cues associated with deceit, and comparing the model performance to improve the classification. The research has chosen a secondary quantitative research approach that has been supported by a deductive and a positivism method, and has used the Kaggle “Real or Fake Job Posting Prediction” dataset with 17,880 job posts. Data were collected by importing the dataset and performing extensive preprocessing, which includes text cleaning, lemmatisation, stop-word removal, filtering rare-word removal and TF-IDF vectorisation. Further, the study has been exploratory data analysis, visualisation, class balancing through SMOTE, and training various models, including Logistic Regression, Naive Bayes, XGBoost, and LSTM. The strongest results have been attained with the Logistic Regression model, which exhibit high recall and F1-scores, especially with respect to the minority fraud group. The study has established that the NLP methods have been fruitfully used in order to extract meaningful linguistic patterns that can be used to identify frauds.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Chikkankod, Arjun UNSPECIFIED
Subjects:	P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management > Human Resource Management H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management > Human Resource Management > Recruitment
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	07 Sep 2026 10:43
Last Modified:	07 Sep 2026 10:43
URI:	https://norma.ncirl.ie/id/eprint/9858

Actions (login required)

View Item