A Comparison of Fake Job Post Prediction Using Different Techniques for Data Mining

Abstract

As a result of current advances in the Internet
of Things or contemporary technology,
individuals are now more likely than ever to
publish career adverts. As a result, anyone is
going to get mindful of the subject of
anticipated fraudulent job advertisements.
Fake job posing prediction is challenging,
similar to many other challenges in this field.
Based on the findings of the examination, a
Random Forest Classifier can be utilized to
determine the legitimacy of a job
advertisement with high accuracy. To test our
proposed theory, we analyzed the
Employment Scam Aegean Dataset
(EMSCAD), which consists of 18,000 records.
A trained classification system can identify a
fake job posting has a certainty of about 98%.
For this classification challenge, a deep neural
network classifier excels. For this deep neural
network classifier, three thick layers were
used. A bogus job advertisement can be
predicted with a classification accuracy of
about 98% by the trained classifier using
DNN. Publicising new job vacancies has
recently become an incredibly regular
problem in today's world because of
improvements in modern innovation and
social communication. Everyone will be
worried about the expectation assignment
from the false job posting as a result. The
same challenges come with fake work
presentation forecasts as they do with other
grouping activities. To predict the task of
determining whether a post is real or fake,
this paper proposed using a variety of
information mining methods and
characterization calculations, such as Support
vector machines, KNN, decision trees
innocent the Probability classification
algorithm, irregular timberland classification
models, multi-facet the perceptron, and
profound brain.