Abstract:
"Amidst the heavy competition in the employment market to secure a job that matches
with one’s qualifications, it is quite a task to seek one’s dream job. Apart from the
unavailability of vacancies, a jobseeker encounters certain other issues in finding
his/her dream job. Technology-related problems occupy major space in present context
since technological advancements have completely taken over the major elements of
employment market: advertising and recruitment. Thus, if either the jobseeker is
unaware of technological aspects involved in employment hunting or the technology
itself causes glitches that will deprive individuals of the opportunity of securing their
dream job.
This study attempts to devise a solution to a technology-based problem in seeking a
better job. The problem exists in the job description of certain most-used job portals. It
is a question that whether the description which is in image format would satisfy jobseeker’s requirements. Thus there is an evident mismatch between these two aspects.
The researcher aimed at minimizing this mismatch by developing a model using
machine learning techniques. The model was designed such that it converts the image
data to text and filters the first ten job descriptions that align with the interest of the
jobseeker. Through Topic Modeling, researcher find the inherent grouping and the
keyword list of the given job description. The algorithms NB, RFC, SGD, and LSVC
provide greater accuracy than the cutoff values. So the job descriptions versus sector
predictions are fed to the Stacking Ensemble Classifier to get the final prediction with
73% of accuracy. Use the Phase Matcher technique to filter the predicted sector jobs
and match the keywords, which are from the topic modelling results. For a given JD,
matching percentages of the keywords are returned eventually. This model is expected
to assist jobseekers in finding a better job that matches with their qualifications that
way they would not miss any opportunity of securing their dream job."
