DEPARTMENT OF COMPUTER ENGINEERING

PROJECT REVIEW II
RESUME RANKING SYSTEM
TEAM MEMBERS
1: Shrushti Sonawane | Roll : 48
2: Nikita Raut | Roll : 49
3: Prajwal Take | Roll : 3
4.Vedant Phatangare | Roll : 14
Under the guidance of
G.B. Kote Sir
 • INTRODUCTION
Contents • PROBLEM STATEMENT
• PURPOSE, SCOPE AND OBJECTIVES
• LITERATURE SURVEY
• SYSTEM OVERVIEW- PROPOSED SYSTEM &
PROPOSED OUTCOME
• SYSTEM REQUIREMENTS
• ARCHITECTURE & DFD
• INDUSTRIAL BENEFITS
• PROJECT PLAN (PWS I)
• REFERENCES
 A proposed machine learning-based system to
INTRODUCTION automate the process of resume screening and

1. Automated resume ranking

1 ranking.

The system utilizes advanced algorithms to

match job descriptions with resumes and rank

2
2. Content-based recommendation
them based on similarity
techniques

3 . Addressing Challenges in resume

screening Traditional methods of finding suitable job

3 applicants face difficulties in efficiently

screening a large number of resumes.
Employers face difficulty in
PROBLEM
identifying the most suitable STATEMENTS
candidates efficiently and
Challenge in finding suitable
accurately.
1 job applicants

Manual screening is time-

consuming, resource-intensive,
prone to human bias and error. Manual resume screening
2 limitations

An automated system is required to

analyze resumes using machine Requirement of Automated
learning techniques, efficiently
screen and rank them for specific
3 System for Resume Filtering

job roles.
PURPOSE

Introduce efficiency to the recruitment process by addressing

the challenges of deciphering CVs and ensuring that
applicants can execute the job before they are hired

"To make it easy for recruiters to efficiently identify the most

qualified candidates for a specific job vacancy."

Using ML techniques (KNN Algo & Cosine Similarity) to develop the

resume ranking system based on content-based recommendation
OBJECTIVE

Main Purpose is To design and develop a

machine learning-based resume Ranking
System .
The system aims to automate the process of screening and ranking
job applicants based on their resumes

Focuses on using content-based recommendation techniques, such as cosine

similarity and KNN algorithm, to identify the most suitable resumes that
closely match the job description.
SCOPE
• Hybrid Recommendation System:
The document suggests building a hybrid recommendation system that combines both
Collaborative Filtering Recommendation Systems and Content-Based Filtering
Systems. This system can utilize the candidate's academic score and MCQ test results
to provide more accurate and personalized resume recommendations.

• Fine-tuning Summarization Technique:

summarization technique used in the "genism" package may result in the loss of
essential information. Therefore, there is a scope to fine-tune this technique to
ensure minimum information loss and improve the quality of the generated
summaries.
Algorithm
KNN
The context of resume ranking, KNN can be used to find the most similar resumes to a
given job description by measuring the distance or similarity between feature
representing the resumes.

Cosine Similarity:
It is a measure used to determine the similarity between two vectors, often used in
recommendation systems. In the resume ranking system, cosine similarity can be
used to calculate the similarity between the vector representations of job
descriptions and resumes. It measures the cosine of the angle between the two
vectors, indicating how similar they are in terms of their content.
Algorithm

TERM FREQUENCY- INVERSE DOCUMENT FREQUENCY (TF-IDF):

TF-IDF (Term Frequency – Inverse Document Frequency) is a handy algorithm that

uses the frequency of words to determine how relevant those words are to a given
documen.
 Literature
Survey
[1] [For Cosine Similarity]SoftSimilarity and Soft Cosine [3] [For KNN] Application of machine learning algorithms
Measure: Similarity of Features in Vector Space Model, to an online recruitment system, in: Proc. International
Computacion y Sistemas 18 (3) (2014) 491–504 Conference on Internet and Web Applications and Services,
By Grigori Sidorov, Alexander Gelbukh, Helena Gomez- 2012, pp. 215–220.
Adorno, David Pinto, . By Evanthia Faliagka, Kostas Ramantas, Athanasios
Tsakalidis, Giannis Tzimas

[2] [For TF-IDF Cal] An automated resume screening [7] [TF-IDF vectorizer] Ranking of text documents using
system using natural language processing and similarity, TF-IDF weighting and association rules mining, in: 2018
Top. Intellig. Comput. Indust. Des. (ICID) 2 (2) (2020) 99– 4th international conference on optimization and
103. applications (ICOA), IEEE, 2018, pp. 1–6.
By Chirag Daryani, Gurneet Singh Chhabra, Harsh Patel, By Siham Jabri, Azzeddine Dahbi, Taoufiq Gadi, Abdelhak
Indrajeet Kaur Chhabra, Ruchi Patel, Bassir,
 Job Desc Processing SYSTEM
To extracts key information, such as
required skills, qualifications, and
experience.from it
1 OVERVIEW
The system aims to automate the
Resume Parsing process of screening and ranking
extracting relevant information from resumes for job vacancies. It
resumes, such as skills, education, and
work experience. 2 utilizes content-based
recommendation techniques to
match job descriptions with
Resume Ranking resumes and rank them based on
calculates the similarity between job descr and similarity
resumes keywords (cosine similarity,KNN
algo) & ranks Resumes based on similarity
with Job Descriptions Requirements and
3
SkillStets .
 PROPOSED OUTCOME
The System will be able to assess each candidate’s resume and assign
a relative rating and score. We Should note that the system has limitations and should not be the
Single criteria for choosing a candidate
Where

Parsing Accuracy:
This indicates that the system Will be aimed to
successfully extract relevant information from resumes,
such as education, work experience, and skills.

Ranking Accuracy:
This means that the system Will try to effectively match
job descriptions with resumes and ranked them based on
their similarity.
 SYSTEM REQUIREMENTS

OS LIBRARIES DBMS
Any Relavant Windows (7, 10, Python 3.7+ with libraries like scikit- SQLite, a lightweight and embedded
11), MacOs, Linux learn, NLTK, and pandas could be used database, could be used for storing and
for implementing the system retrieving resume and job description
data.

RAM DISK SPACE PROCESSOR

4GB or above 256 GB or above (ssd prefferd) any latest generation chip with
Basic Computational power
ARCHITECTURE AND DFD

DISK SPACE PROCESSOR

Fig. Framework of
the resume screening
and ranking system.
 Industrial Benefits
Improved Hiring Decisions:
By using content-based suggestion and cosine similarity, the system can match job
descriptions with resumes and recommend the top candidates. This ensures that recruiters can make
more informed hiring decisions based on the skills and qualifications of the applicants.

Efficient Screening:
The automated resume ranking system offers a more efficient screening process for recruiters. It
helps them quickly identify the most qualified candidates for a specific job vacancy, saving time and
resources.
Scalability:
The proposed system can handle large quantities of resumes and job descriptions, making it suitable
for organizations with high recruitment needs. It can efficiently process and rank a vast number of
resumes, enabling recruiters to find the right candidates from a large pool of applicants.
REFERENCES
[4] Evanthia Faliagka, Kostas Ramantas, Athanasios Tsakalidis,
[1] Pradeep Kumar Roy, Sarabjeet Singh
Giannis Tzimas, Application of machine learning algorithms to an
Chowdhary, Rocky Bhatia, A Machine Learning
online recruitment system, in: Proc. International Conference on
approach for automation of Resume
Internet and Web Applications and Services, 2012, pp. 215–220.
Recommendation system, Procedia Comp. Sci.
167 (2020) 2318–2327.
[5] Shaha T. Al-Otaibi, Mourad Ykhlef, A survey of job
recommender systems, Int. J. Phys. Sci. 7 (29) (2012)
[2] Jyothis Joseph, Jaimy Sunny, R Raveena, 5127–5142.
BlessyElzaByju, KC Laya, Resume Analyser:
[6] Adem Golec, Esra Kahya, A fuzzy model for
Automated Resume Ranking Software, Int. J. Res.
competency-based employee evaluation and selection,
Appl. Sci. Eng. Tech. (IJRASET) 8 (7) (2020)
Comput. Ind. Eng. 52 (1) (2007) 143–161.
896–899.

[7] Paul Resnick, Hal R. Varian, Recommender

[3] Aseel B. Kmail, Mohammed Maree,
systems, Commun. ACM 40 (3) (1997) 56–58.
Mohammed Belkhatir, Saadat M. Alhashmi, An
automatic online recruitment system based on
exploiting multiple semantic resources and [8] Evanthia Faliagka, Kostas Ramantas, Athanasios Tsakalidis, Giannis
concept-relatedness measures, in: 2015 IEEE, Tzimas, Application of machine learning algorithms to an online recruitment
2015, pp. 620–627. system, International Conference on Internet and Web Applications and
Services, 2012, pp. 215–220.
THANK YOU