ANALYSIS OF RESUME USING MACHINE

LEARNING

Ullas U Dhanush M R Vikas Jain

Department of Information Department of Information Department of Information
Science and Engineering Science and Engineering Science and Engineering
Vidyavardhaka College of Engineering Vidyavardhaka College of Engineering Vidyavardhaka College of Engineering
Mysuru, India Mysuru, India Mysuru, India
[email protected] [email protected] [email protected]

Prof.Megha.V
Department of Information
Science and Engineering
Vidyavardhaka College of Engineering
Mysuru, India
[email protected]

Abstract
This abstract focuses at the transformative and dynamic character of employment in
recent markets. Rapid technological developments and societal shifts,Preferences, as
well as global economic factors, have changed the labour market, posing both
obstacles and possibilities for both job seekers and employers. This research looks at
significant trends in modern employment, such as the increase of remote work, the
relevance of digital skills, the advent of the gig economy, and the rising emphasis on
workplace sustainability and diversity. Furthermore, it investigates the influence of
these developments on job searchers, businesses, and politicians, emphasising the
importance of adaptation, lifelong learning, and forward-thinking legislation in
order to navigate and prosper in an ever-changing labour market. Understanding the
dynamics of contemporary markets and their implications for employment is
becoming increasingly important as the world changes. Understanding the dynamics
of recent markets and their implications for employment is critical for individuals,
organisations, and governments seeking to capitalise on the potential of this new era.

Keywords: Authentication, Natural Language processing, Web Technology.

Introduction
In tracing the history of the job market, it becomes evident that the landscape has
continually evolved, shaped by historical events, economic shifts, and technological
advancements. The Industrial Revolution marked a significant turning point,
introducing mass production and altering the nature of work. Subsequent decades saw
the rise of service-oriented economies, globalization, and the advent of the digital age,
each leaving an indelible mark on employment patterns. As we delve into the present,
an analysis of the job market reveals a complex interplay of factors. Rapid
technological advancements, changing societal preferences, and global economic shifts
have transformed the very fabric of the labor market. The subsequent paragraphs
provide an in-depth examination of this transformation.

Job market analysis underscores the contemporary challenges and opportunities that
characterize the employment landscape. Technological disruptions and changing
consumer preferences have spurred a shift in job requirements, with an increasing
emphasis on digital skills and adaptability. The rise of remote work, the gig economy,
and a growing awareness of workplace sustainability and diversity further redefine
traditional notions of employment. Against this backdrop, it is crucial to examine the
unemployment rate, a metric that reflects the delicate equilibrium between job seekers
and available opportunities. Understanding the dynamics of unemployment provides
insights into the broader economic health and societal well-being.

In addressing the transformative and dynamic nature of employment, it is imperative to

acknowledge the multifaceted impacts of global economic factors, preferences, and
technological advancements. The penultimate paragraph delves into the significant
trends shaping modern employment. Remote work has gained prominence, altering
traditional workplace structures and creating new possibilities. The demand for digital
skills has surged, reshaping job requirements and creating opportunities for those with
technological proficiency. The gig economy has introduced a more flexible approach to
work, challenging conventional employment norms. Moreover, the rising emphasis on
workplace sustainability and diversity reflects a societal shift towards more inclusive
and environmentally conscious workplaces.

This research recognizes that these transformative shifts in the labor market pose
challenges and opportunities for job seekers, businesses, and policymakers alike. The
adaptability of individuals, the commitment to lifelong learning, and the formulation of
forward-thinking legislation become essential strategies to navigate and thrive in an
ever-changing employment landscape. As the world undergoes significant changes,
understanding the dynamics of recent markets and their implications for employment is
critical for individuals, organizations, and governments aiming to capitalize on the
potential of this new era. Keywords: Authentication, Natural Language Processing,
Web Technology, Wireless Fidelity (Wi-Fi).

The survey paper provides insights into the dynamic nature of modern employment,
examining historical shifts and current trends. It explores challenges and opportunities
posed by technology, societal preferences, and global economic factors. The
methodology section outlines the development of a sophisticated resume-based job
matching system using Natural Language Processing and machine learning. The
conclusion underscores the importance of adaptation and lifelong learning in navigating
the evolving job market. References substantiate the research, contributing to a
comprehensive understanding of contemporary employment dynamics.

Literature Survey
In order to improve the effectiveness of the hiring process, Chirag Daryani et al. [1] present
an automated resume screening system that makes use of similarity measures and Natural
Language Processing (NLP). The system creates condensed resumes for examination by
using natural language processing (NLP) techniques to extract pertinent information from
unstructured resumes. A vectorization model and cosine similarity are used to rank
candidates according to how well they match the job description. Information extraction,
feature selection, vectorization, and similarity computation are among the architectural
components. In addition to exploring the features of the current system, the study makes
suggestions for future developments that could enhance it further, like integrating semantic
similarity metrics and social networking data. This study advances the continuous
development of automated candidate screening systems by highlighting the fusion of NLP
and similarity measures for more effective and nuanced recruitment processes.

Irfan Ali et al.[2] introduced fusion of NLP and similarity measures for more effective and
nuanced recruitment processes.The creation of an automated Resume Classification System
(RCS) with the use of machine learning (ML) and natural language processing (NLP)
techniques is covered in the document. Based on their content, resumes are supposed to be
categorised by the system into several job categories. The work focuses on resume
preprocessing, feature extraction, and representation utilising Word-to-Vec, N-grams, and
Bag of Words. The effectiveness of many machine learning classifiers, such as Support
Vector Machines (SVM), Naïve Bayes, K-Nearest Neighbours, and Logistic Regression, is
assessed in the resume classification task. The study concludes that with over 98%
accuracy, SVM classifiers work the best. The suggested solution can assist recruiters in
overcoming obstacles like manual screening, matching resumes to job specifications, and
cost management by automating the tiresome resume selection process. Additionally, the
study also highlights the importance of finding a suitable and standardized dataset for
training the ML models. Overall, the system shows promising results and can be deployed
in real-time applications to expedite the resume screening process.
The research of Kemas M Lhaksmana et al.[3], undertaken at Telkom University's
School of Computing, dives into the use of TF-IDF and Word Embedding for
automating job interview grading, with the overriding goal of decreasing biases and
human mistakes in the selection process. In order to solve the issues that firms have in
talent acquisition, the study evaluates the performance of TF-IDF and Word Embedding
using an Artificial Neural Network (ANN) classifier. The dataset, obtained from PT
Telkom Indonesia, contains interview transcripts from 56 candidates, which were
graded using nine different criteria. The significant finding demonstrates TF-IDF's
superiority, with an average accuracy of 85.22% compared to Word Embedding's
74.88%. The document emphasises not only the usefulness of TF-IDF, but also the
possibility of increased accuracy through the development of the model. Beyond the
local context, the study highlights the larger implications of automation in job interview
grading, highlighting the critical significance of feature extraction approaches in
assuring accurate text categorization. This study adds to the continuous advancement of
AI applications in recruiting procedures, opening the way for more efficient, unbiased,
and accurate talent selection methods.

Minyar Sassi Hidri’s et al[4] document delineates the development and evaluation of a
bi-directional recommendation system utilizing natural language processing (NLP) to
align job offers with candidate profiles and vice versa. Employing machine learning
methodologies, the system's performance is rigorously assessed through decision
support metrics such as accuracy, precision, recall, and F1-score. The primary objective
is to enhance the efficiency and effectiveness of the recruitment process for both job
seekers and employers, with the potential to mitigate frictional unemployment by
improving information accessibility and reducing time and opportunity costs associated
with job search and recruiting. The study suggests future enhancements, including
considerations for relevant experience dates and improved talent identification in job
offers. Positioned within a broader framework for enhancing computerized recruiting
processes, the text underscores the significance of personalized reciprocal referral
systems to address frictional unemployment. The key contribution lies in the creation of
a recommendation system facilitating optimal matches for job seekers and recruiters,
supported by an in-depth exploration of NLP techniques for text pre-processing, vector
representation, and similarity measurement, all assessed through robust decision support
metrics.
Suleiman Ali Alsaif et al.[5]This article addresses the setup and assessment of the JRS
(Job Recommender System) system for job recommendations. In order to suggest
positions to job seekers based on their resumes and talents, the system uses Natural
Language Processing (NLP) to gauge how comparable job titles and job descriptions
are. The algorithm determines similarity using the Jaccard coefficient and cosine
distance in conjunction with content-based filtering. Machine learning decision support
metrics including accuracy, precision, recall, and F1-score are used to assess the
system's effectiveness. A dataset of job offers that were taken from Saudi Arabia's
major cities was used to test the system. The usage of stop words filtering,
lemmatization, and tokenization in data pre-processing is also mentioned in the
document. The purpose of the suggested method is to offer tailored employment
recommendations to ob seekers based on their skills and the requirements of job
listings.
Mantas Lukauskas [6] addresses about how they implemented this automated job
profiling using NLP Model. The usage of Python modules like Playwright,
BeautifulSoup, and Selenium is one of the techniques for data extraction from
unstructured texts that is examined in this paper. The use of the regex method for data
extraction, which produced positive results for extracting data from Lithuanian
employment adverts, is also included in the study. The study also explores the
processing of unstructured data from job advertisements using feature extraction
techniques like TF-IDF and sentence transformers (BERT model). By evaluating job
ads and utilizing natural language processing (NLP) and clustering approaches to
segment them into profiles, this research work aims to improve understanding of skills
need. Examine how the dynamics of particular abilities advertised in job ads vary over
time by tracking changes in frequency.To find the best clustering and dimensionality
reduction strategies for this work, compare various approaches. Group job postings
according to the skill requirements to determine the primary job profiles and categories
in the Lithuanian labor market.By using a data-driven examination of job adverts,
provide insights into how the skill requirements in the Lithuanian labor market are
changing.Provide an automated online job posting analysis approach to policymakers
and other stakeholders to evaluate changes in the demand for specific job skills in real
time.In conclusion, the primary objective is to enhance comprehension of the skills
demand by profile-segmenting job ads by machine learning methods such as clustering,
dimensionality reduction, and natural language processing. This seeks to assist
employers, job seekers, and educators and policymakers respond to evolving skill
needs.
In this paper Ganesh Kumar explains [7] about how a resume can be parsed using
different text mining and NLP techniques to get the particular key word. The document
discusses a novel job portal with a resume evaluation system based on text mining and
NLP techniques. It proposes a system to match job seekers with recruiters using web
crawling, content mining, and natural language processing. The system aims to simplify
the recruitment process and provide automatic recommendations to job seekers based
on their skills. It also addresses the issue of resume replication in existing job portals.
The main problem addressed in existing job portals, according to the document, is the
excessive replication or lack of information on job openings, leading to confusion and
inefficiency for job seekers. Additionally, there is a gap between the skills required by
employers and the skills possessed by job seekers, making it challenging to match the
right candidates with the right job opportunities. The system provides recommendations
to job seekers based on their skills through the use of a cooperative filtering algorithm.
This algorithm matches the skills of individual job seekers with the requirements of job
postings, allowing for personalized and relevant recommendations. Additionally, the
system leverages natural language processing techniques to analyze the skills and
qualifications of job seekers, ensuring that the recommendations are tailored to their
specific abilities and expertise. By employing these methods, the system aims to
enhance the job search experience for candidates by presenting them with job
opportunities that align with their skill sets.
The article [8] presents a comprehensive approach to developing a job recommendation
system that goes beyond traditional methods. By utilizing advanced NLP techniques
such as Word2Vec and Gensim, the system aims to improve recommendation quality by
addressing semantic nuances. The article also suggests integrating machine learning
algorithms and creating an Android application to accommodate regional language job
searches, making the system more accessible and user-friendly. Furthermore, it
anticipates the potential impact of these advancements on streamlining the job search
process and providing users with more tailored and relevant job opportunities. In
conclusion, the article demonstrates the ongoing evolution of job recommendation
systems, continually striving to enhance user experience through advanced NLP
technologies.

Methodology

The Job Scouter project

employs a robust client-
server architecture,
utilizing cutting-edge
technologies to create a
transparent and efficient
platform for delivering
information.

To navigate the challenges encountered during the development of our full-stack AI/ML
application, our strategic approach unfolded as follows. Firstly, meticulous
consideration was given to selecting reliable data sources, streamlining the time-
consuming process of acquiring pertinent and high-quality data. A key focus was placed
on refining the technique for removing stop words from resumes to enhance accuracy
without compromising crucial information. Overcoming the intricacies of LinkedIn's
API involved surmounting hurdles related to authentication, ensuring secure data
extraction. Following data retrieval, a sophisticated procedure for reading and parsing
API data was implemented to present information in an accessible format, guaranteeing
a seamless user experience. Effectively bridging the language gap between Python
scripts and Node.js allowed for the execution of Python scripts within the Node.js
runtime. Lastly, persistent efforts were directed towards curating a comprehensive
dataset to address the challenge of obtaining sufficient and diverse data for model
training, a crucial aspect for the robust training of our machine learning models.

Conclusion
In light of the limitations inherent in existing keyword-based search methods, the
challenge of effectively aligning skilled job seekers with suitable employment
opportunities has become more apparent in today's dynamic labor market. The Job
Scouters initiative addresses this challenge through the development of an innovative
resume-based job matching system. Our goal is to enhance the precision of job-seeker
and position matching by leveraging sophisticated Natural Language Processing (NLP)
and machine learning techniques. The core strength of our system lies in intelligent
resume analysis, utilizing powerful NLP algorithms to comprehend nuanced details
beyond traditional keyword matching. This information is then dynamically cross-
referenced with a diverse array of job advertisements from multiple platforms, resulting
in a personalized job suggestion system. Recognizing the evolving dynamics of the
modern market and their implications for employment is increasingly crucial as the
global landscape transforms. Understanding these dynamics is essential for individuals,
businesses, and governments seeking to capitalize on the promises of this new era.

REFERENCES:
1. Chirag Daryani,aGurneet Singh Chhabra,Harsh Patel, Indrajeet Kaur Chhabra, Ruchi
Patel.”Resume Classification System using Natural Language Processing and
Machine Learning Techniques”,accepted 26 November 2020.

2. Irfan Ali,, Nimra Mughal, Zahid Hussain Khand,, Javed Ahmed,Ghulam

Mujtaba.”Resume Classification System using Natural Language”,accepted on
05.05.2021.

3. Annalisa Wahyu Romadon,Kemas M Lhaksmana, Analyzing TF-IDF and Word

Embedding for Implementing Automation in Job Interview Grading(2020) Accepted
July 15 2020.

4. Suleiman Ali Alsaif, Minyar Sassi Hidri,NLP-Based Bi-Directional

Recommendation System: Recommending Jobs to Job Seekers and Resumes to
Recruiters(2022) Accepted November 24 2022.

5. Alsaif, S.A.; Sassi Hidri, M.; Eleraky, H.A.; Ferjani, I.; Amami, R Learning-Based
 Matched Representation System for Jobn Accepted: 7 November 2022.

6. Enhancing Skills Demand Understanding through Job Ad Segmentation Using NLP

and Clustering Techniques by Mantas Lukauskas 1,*ORCID,Viktorija Šarkauskaitė
2,Vaida Pilinkienė 3ORCID,Alina Stundžienė 3ORCID,Andrius Grybauskas 3
andJurgita Bruneckienė 3

7. A Novel Job Portal With Resume Evaluation System Based An Text Miningand Nlp
Techniques S. Ganesh Kumar1, Ritvik Jain2, Ashish Singh Chauhan3

8. : Job Recommendation System Using NLP By Sneha Kumari, Gittika Agarwal, Dr.
Sheshappa S.N, Mr. Vijaykumara.Y.M., and Prof. Mr. Byre Gowda B.K.

Prof.Megha.V Ullas.U Vikas Jain Dhanush M R Prof.Ravi.Kumar