Employment Recommendation System Using Content-Based Filtering Approach and Fuzzy Logic
Date
2020-08-19
Authors
Amani Basem Mohammed Haboub
أماني باسم محمد حبوب
Journal Title
Journal ISSN
Volume Title
Publisher
Al-Quds University
Abstract
Nowadays, there is a gap between the requirements of the labor market and the scientific
specializations, which are available in universities. The existence of similar disciplines with
different names in universities leads to a large number of specializations, especially for a
specific number of jobs. These things we mentioned before make it difficult to find and
match the most suitable specialization, which based on the educational outcomes for the
available jobs in the labor market.
The importance of the study reveals in matching Automatic job with an available job
positions is a big problem for organizations and applicants. There are many techniques and
algorithms to help job seekers find the right job. Some are traditional algorithms while others
have been found recently, also there are a large number of hybrid algorithms that are a
combination of many algorithms. All these algorithms aim to find the best job for the
candidate and to make recruitment processes faster, more accurate and transparent.
The literature review focused on how we can help job seekers find the most suitable jobs
using techniques such as analysis, extraction, and web crawling. The problem with current
method is that it calculates the similarity between job applications (job profile) and CV (user
profile, this happens without measuring whether the recommended candidate is satisfying or
meets the employer's goals or needs.
The main question of this work (problem statement) is how to achieve the best matching
between candidates and offered jobs. So, we suggest a hybrid system based on NLP tasks,
BabelNet dictionary, content base filtering approach and finally fuzzy logic, all that allow
system users (i.e., applicants and employers) to search for suitable jobs with matching them
with job offers, which leads to best presentation for each candidate. We also suggest building
a database contains different specialties and appropriate jobs, according to the opinion of
stakeholders, such as academics who teach that specialization and the opinion of some
workers in the labor market.
Our main contribution is proposing a model to manage the recruitment process by evaluating
the applicants' CVs, which contents weighting match to the offered/posted (open jobs)
requirements by recognizing the employers' needs.
In order to make our finding approach closer to reality, transparent and to reduce effort and
time applicants and employers spend, we take into account the opinions of academics
working in Palestinian universities in different disciplines and their relationship to the labor
market.
Also, the study provides an opportunity for applicants, who were not accepted when the
education section was evaluated from 100%, or getting a low match ratio to be acceptable,
and has a good chance to employ. The expert knowledge base we created effects on results by
giving a wider chance to those who had a low employing chance to be accepted, also the
study shows that the employee opinion effected in the candidate list.
Finally, as appears from candidate lists, the system is well effective in extracting lists
according to the proportion of matching required for each section in the employment
announcement, and in terms of accuracy. The results are satisfactory, according to seekers
qualifications, there are very close matching rates in some cases and spaced out in other
cases. Also, it is easy to use through the simple user interface and flexibility in dealing with
by consistent with the requirements of the employer expressed by a certain percentage ratioالوظيفة الشاغرة مع المتقدمين لها، لاستخراج أفضل النتائج.
في الوقت الحاضر، هناك فجوة بين متطلبات سوق العمل والتخصصات العلمية المتاحة في الجامعات. وقد أدى وجود تخصصات مماثلة بأسماء مختلفة في الجامعات، إلى وجود عدد كبير من التخصصات لعدد محدد من الوظائف. وذلك جعل من الصعب العثور على التخصص الأنسب للوظائف المتاحة في سوق العمل. تعد المطابقة التلقائية للوظائف والمرشحين، مشكلة كبيرة للشركات والمتقدمين للعمل معا. هناك العديد من التقنيات والخوارزميات لمساعدة الباحثين عن عمل في العثور على الوظيفة المناسبة، تهدف كل هذه الخوارزميات إلى إيجاد أفضل وظيفة للمرشح الانسب، وجعل عمليات التوظيف أسرع وأكثر دقة content-based وشفافية. في هذه الأطروحة، نقترح نهجًا يستخدم طريقة التوصية القائمة على المحتوى لمطابقة BabelNet وقاموس Fuzzy logic جنبًا إلى جنب مع المنطق الضبابي filtering approach
في الوقت الحاضر، هناك فجوة بين متطلبات سوق العمل والتخصصات العلمية المتاحة في الجامعات. وقد أدى وجود تخصصات مماثلة بأسماء مختلفة في الجامعات، إلى وجود عدد كبير من التخصصات لعدد محدد من الوظائف. وذلك جعل من الصعب العثور على التخصص الأنسب للوظائف المتاحة في سوق العمل. تعد المطابقة التلقائية للوظائف والمرشحين، مشكلة كبيرة للشركات والمتقدمين للعمل معا. هناك العديد من التقنيات والخوارزميات لمساعدة الباحثين عن عمل في العثور على الوظيفة المناسبة، تهدف كل هذه الخوارزميات إلى إيجاد أفضل وظيفة للمرشح الانسب، وجعل عمليات التوظيف أسرع وأكثر دقة content-based وشفافية. في هذه الأطروحة، نقترح نهجًا يستخدم طريقة التوصية القائمة على المحتوى لمطابقة BabelNet وقاموس Fuzzy logic جنبًا إلى جنب مع المنطق الضبابي filtering approach