Here, I am creating a model that discriminates between legit and illegitimate job posts.
I use metrics such as logo, salary range, and questions, as key indicators on whether or not a job is legit. Other things I am looking at here include trying to answer these questions; which jobs are likely fake, and what variables are key predictors of this likelihood. Another question that I attempt to answer here concerns the best threshold to operate in such that most fake jobs are purged, while keeping false positive at a tolerable minimal.