Kisko V. Apeles

Kisko V. Apeles

Master of Science in Data Science - Asian Institute of Management

JOB OR SCAM: Identifying Fraudulent Job Ads using Machine Learning

1 minute read

by Kisko Apeles, Ronald Dela Cruz, Nisarg Nigam, Ralph Palomaria, M Hwang, and Maynard Vargas

png

Last October 2019, the Department of Labor and Employment (DOLE) publicly warned all job seekers on the proliferation of fake local and overseas job listings in social media.

These fraudulent job listings aim to:

  • Steal personal information
  • Unfairly acquire money
  • Stain organization reputations

This doesn’t just happen in the Philippines. In the US, financial losses can amount up to 1.25 Million USD.

Hence, our group aimed to identify these fraudulent job listings.

Using the Employment Scam Aegean Dataset which had ~ 18,000 job listings, our group leveraged on different natural language processing techniques and machine learning models to identify these fraudulent job listings.

Our best model achieved an 80% accuracy (F1-score) and we discovered that fraudulent job listings have:

  • Entry-level position indication
  • Short company profiles
  • Missing information
  • Spelling errors

Our model and its results can be used for the following benefits:

  • Prevention of employment scams
  • Incorporation of model in online job websites
  • Cautioning individuals

With that, we hope we can all help our fellow job seekers avoid being scammed and have a good job!

References:

  • Department of Labor and Employment. (Oct 13, 2019). Public warned on fake jobs. Retrieved Sept 01, 2020 from https://www.dole.gov.ph/news/public-warned-on-fake-jobs/
  • Crane, C. (Jan 28, 2020). Fake Jobs: Cybercriminals Prey on Job Seekers via Fake Job Postings. Retrieved Sept 01, 2020 from https://securityboulevard.com/2020/01/fake-jobs-cybercriminals-prey-on-job-seekers-via-fake-job-postings/

You May Also Enjoy