Resume Parsing System for Software Job Openings

Ms. Gunjan Sharma Mr. Tushar Kumar Tailor

MIT Art, Design and Technology, Rajbaug, MIT Art, Design and Technology, Rajbaug,
Loni Kalbhor, 412201 Loni Kalbhor, 412201
+91-8355982670 +91-7357170421
gunjan2607sharma@gmail.com tusharkumartailor@gmail.com

ABSTRACT
The hiring units of organizations, especially huge enterprises, parser system can be used to extract key information such as work
often find the recruitment process very tedious, for lack of on- experience, education, skills, and contact details from a
point shortlisting of eligible candidates for relevant job openings. candidate's resume. This can be useful for recruiters and hiring
Considerable time and human error can be reduced if a good managers who need to quickly evaluate a large number of resumes
recommendation system for candidates is used. This paper focuses and identify the most qualified candidates.
on methods to extract the required information about candidates
Here's something to look very important, no parser system is
without having to go through each resume manually. The Resume
perfect, and there may be errors or omissions in the extracted data.
Parser System (RPS)- uses Natural Processing Language
Additionally, different parser systems may use different
(NLP)and statistical programming. The text in the resume is
algorithms and techniques, which can affect the accuracy and
parsed and tagged based on skill sets relevant to job profiles in the
efficiency of the parsing process. Therefore, it is important to
software industry. The aim is to study and implement a parsing
carefully evaluate and test different parser systems before
system that picks resumes with the most relevant based on the
selecting one for use in recruitment processes.
feature analysis.

Keywords 2.2 Candidate Skillset Database

When there is a very concrete understanding of what the system
Natural Language Processing; Text Analysis; Text
means and the parsing of the resume. We try to understand the
Summarization; Recommendation System.
further module of it which plays a very integral part in the process
of building the system. To create a comprehensive candidate
skillset database, recruiters can use a variety of tools and
1. INTRODUCTION techniques such as job analysis, candidate screening, and skills
A resume parsing system designed to help students analyze their assessment tests. These methods can help identify the specific
skill sets and identify areas for improvement. The system utilizes skills and competencies required for a given job position and can
natural language processing techniques to extract relevant help evaluate the qualifications of job candidates to determine
information from resumes and store it in a structured format. By their suitability for the job.
analyzing old resumes of people that are working in the industry, It's important to keep candidate skillset databases up to date to
the system provides suggestions on potential target companies for ensure that the information is accurate and relevant. This can
students, as well as the likelihood of being hired by those involve regularly updating candidate profiles, conducting skills
companies. In addition, the system also suggests areas for assessments and performance reviews, and providing ongoing
improvement to increase a student's chances of getting hired by training and development opportunities to employees.
their desired company. The overall goal is to enable students to
develop the skills they need to succeed in their future careers.
2.3 Resume Ranking Algorithm
2. Resume Ranking System This subsection includes the set of rules or procedures used to
A resume ranking system is a software application that uses evaluate resumes and determine their relevance and suitability for
machine learning algorithms to analyze and evaluate resumes a particular job opening. The algorithm can be based on a variety
based on specific criteria. The system assigns a score or ranks to of factors such as the skills and qualifications listed on the
each resume, which helps recruiters and hiring managers identify resume, the candidate's work experience, educational background,
the most qualified candidates for a particular job. The criteria used and other relevant information.
to rank resumes may include the relevance of the candidate's skills Some examples of Ranking Algorithms:
and experience to the job requirements, the quality and readability
of the resume, and other factors such as education, certifications, 2.3.1 Keyword Matching:
and language proficiency. By automating the resume ranking This algorithm scans resumes for keywords related to job
process, the system helps recruiters save time and improve the requirements and ranks them based on the frequency and
efficiency of the recruitment process. This includes the following relevance of those keywords. It generates a list of keywords or
subsections in its processes. phrases related to the job opening, such as required skills,
experience, or education. It then scans each resume for the
presence of these keywords and assigns scores based on how
2.1 Parser System many times each keyword appears in the resume, as well as the
The tool extracts the relevant information from a given input and context and relevance of each appearance.
transforms it into a structured format. In the context of resumes, a
2.3.2 Artificial Neural Networks: 2.5 Database Management Module
This algorithm uses machine learning techniques to analyze
It is a crucial component of a resume parsing system. It is
resumes and learn which attributes are most important for a
responsible for creating and managing a database of candidate
particular job. The algorithm then assigns scores to resumes based
resumes that can be easily searched and queried to identify the
on how closely they match those attributes.
most qualified candidates for a job opening. The module typically
2.3.3 Latent Semantic Analysis: includes functions for importing and storing resume data in a
This algorithm analyses the content of resumes and job structured format, such as XML or JSON. The data is then
descriptions to identify patterns and relationships between indexed and optimized for fast searching and retrieval using a
different terms and phrases. It then uses this analysis to assign search engine or database management system.To ensure the
scores to resumes based on their semantic similarity to the job accuracy and completeness of the database, the module may
requirements. include validation and verification functions to check for errors
2.3.4 Random Forest: and inconsistencies in the resume data. It may also include data
This algorithm uses decision trees to analyse resumes and assign cleaning functions to remove irrelevant or redundant data and to
scores based on the attributes that are most important for a standardize the data format for consistency and ease of analysis.
particular job. The algorithm can be trained on a set of historical The database management module also plays a key role in
data to improve its accuracy and performance. ensuring data privacy and security. It may include functions for
2.3.5 Natural Language Processing: access control, encryption, and data backup to protect the
This algorithm uses advanced text analysis techniques to identify sensitive information contained in candidate resumes.
and extract relevant information from resume. some common
NLP techniques used in resume parsing systems include part-of-
speech tagging, named entity recognition, and sentiment analysis. 2.5 User Interface
The use of NLP in resume parsing systems for software job
the user interface plays a crucial role in facilitating the recruitment
profiles has significantly streamlined the recruitment process for
process. The user interface allows the user to easily navigate the
software jobs, allowing recruiters and hiring managers to quickly
system, input job requirements, and access parsed resume data. It
identify and screen top candidates
also includes features such as filtering options and a dashboard for
2.4 Pre-Processing Module tracking candidates. It should also display parsed data in an
Pre-processing tasks help to prepare the resume data for further organized and easy-to-understand manner, such as highlighting
analysis and processing in the resume parsing system. By cleaning key skills, work experience, and education. In this system, we
and extracting relevant information, converting text to a particularly have very important insights delivered like the resume
standardized format, and identifying key entities, the system can score, and skill assessment, categorizing them into intermediate,
more accurately evaluate candidate resumes and identify the most beginner and experienced levels, which helps in the self-analysis
qualified candidates for a particular job opening. of the candidate.

3. Methodology
Table 1. Tasks in a Pre-Processing System
The methodology for a resume parsing system is a complex and
iterative process that involves careful data collection, pre-
Pre-Processing
Description processing, feature extraction, algorithm development, and
Task
evaluation. By following a systematic and rigorous approach,
Removal of irrelevant and redundant data, developers can ensure that the resulting system is accurate,
Data Cleaning
such as headers, footers, and graphics. effective, and able to identify the most qualified candidates for a
Extraction of relevant text from the resume specific job opening.
Text Extraction file, such as contact information, education,
work experience, and skills.
3.1 Research design and approach
This study will use a quantitative research design with a
Breaking down the text into individual
comparative approach to evaluate the performance of different
Tokenization tokens or words, which can be analyzed and
resumes. This involves identifying the specific needs and
processed further.
challenges faced by candidates in understanding their priorities in
Elimination of commonly used words that do
Stop words the software industry, conducting a literature review, defining
not add meaning to the resume, such as "a",
Removal system requirements, developing, and testing prototypes,
"an", "the", and "in".
collecting and analyzing data, refining the system based on
Reduction of words to their root form to
feedback and data, and validating the system's effectiveness in
simplify analysis and comparison, such as
Lemmatization real-world scenarios. This approach prioritizes user needs and
converting "running", "runs", and "ran" to
feedback while leveraging advancements in natural language
"run".
processing and machine learning.
Conversion of text to a standardized format
Normalization to ensure consistency and accuracy in
analysis and processing. 3.2 Data Collection and analysis methods
Identification of named entities, such as The process for data collection and analysis methods in a resume
Entity organizations, people, and locations, which parsing system typically involves several steps to ensure that the
Recognition can provide context and insight into the system is effective and accurate. Here are some general steps to
candidate's experience and skills. follow:
 1. Identify the data sources: since the module is limited to company, the sample should include resumes from
the software job profiles, for us in this case the data candidates with experience in software development.
source was our department. We could understand and 4. Balance: The sample of resumes should be balanced in
get a better sense of the model of what we are building terms of the number of resumes from each job category,
up. industry, and level of experience. This will ensure that the
2. Collect the data: Once the data sources have been system is not biased toward any particular category or
identified, the next step is to collect the data. We group.
achieved this through the outcome of the survey that we
had by spreading the google form. The collection of
data can be majorly counted with the web scrapping that 3.4 Variables and measurements
we did through various authorized websites. In a resume parsing system, variables and measurements play a
3. Pre-process the data: After the data has been collected, crucial role in ensuring the accuracy and effectiveness of the
it needs to be pre-processed to ensure that it is clean and system. Variables are the factors that are used to evaluate and
formatted correctly. This can include removing score a resume, such as skills, education, work experience, and
irrelevant information, standardizing formats, and certifications. These variables are typically extracted from the
correcting errors. resume during the pre-processing stage and are used to develop
4. Analyze the data: Once the data has been pre-processed, algorithms that can accurately identify and rank the most qualified
it can be analyzed using various techniques such as candidates for a specific job opening.
natural language processing (NLP) and machine Measurements, on the other hand, are the criteria that are used to
learning algorithms to identify patterns and extract evaluate the performance of the resume parsing system. These
relevant information. measurements can include metrics such as precision, recall, and
5. Train the system: The analyzed data then had to be F1 score, which are used to evaluate the accuracy and
trained and moulded according to the needs of the effectiveness of the system in identifying qualified candidates.
system. This involves feeding the system with large Other measurements might include the time it takes for the system
amounts of data to teach it how to recognize and extract to parse a resume, the accuracy of the extracted information, or
the relevant information. the system's ability to handle different types of resume formats.
6. Test the system: After the system has been trained, it
needs to be tested to ensure that it is effective and 4. Discussions
accurate. This involves using a set of test data to The research findings suggest that the resume parsing system
evaluate the system's performance and identify any offers an efficient and accurate way for students to analyze their
issues or areas for improvement. skill sets and identify areas for improvement. By suggesting
7. Refine the system: Based on the results of the testing potential target companies and providing information on how to
phase, the system can be refined and improved to ensure improve their chances of being hired, the system has the potential
that it is as accurate and effective as possible. to significantly improve the educational process by enabling
students to develop the skills they need to succeed in their future
careers. The analysis of old resumes of people that are working in
3.3 Sample size and selection criteria the industry adds a valuable perspective to the system's feedback,
3.3.1 Sample size providing students with insights into the expectations and
The sample size for a resume parsing system will depend on the requirements of employers.
size and diversity of the job market being analyzed. Typically, a
larger sample size will provide more accurate and representative 4.1 Key Findings
results. However, collecting and analyzing large amounts of data The key findings of a resume parsing system for software jobs can
can be time-consuming and resource-intensive. A minimum vary depending on the specific implementation and evaluation
sample size of 100-200 resumes which is specific to the software methods used. Some common key findings that may emerge from
industry is often recommended to ensure that the system is trained such a study include:
on a diverse range of software job profiles.
4.1.1 The importance of keyword matching:
One key finding may be that the system's ability to accurately
3.3.2 Selection Criteria identify relevant keywords and match them to a job description is
The selection criteria for the sample of resumes used to train a crucial to its overall effectiveness. This may be particularly true
resume parsing system will depend on the job market being for technical roles in software development, where specific
analyzed and the specific requirements of the system. Some programming languages, tools, and technologies may be required.
common selection criteria include: 4.1.2 The impact of resume format:
1. Diversity: The sample of resumes should be diverse in Another key finding may be that certain resume formats are more
terms of job titles, industries, and levels of experience. This effective than others in conveying relevant information and
will ensure that the system is able to recognize and extract enabling the parsing system to accurately extract key variables.
relevant information from a wide range of job profiles. For example, a well-structured resume with clear section headings
2. Quality: The resumes in the sample should be of high and bullet points may be easier for the system to parse than a
quality, with clear and well-structured content. This will resume with a more free-form structure.
ensure that the system is able to accurately extract the
relevant information. 4.1.3 The importance of data quality:
3. Relevance: The resumes in the sample should be relevant A key finding may be that the quality of the data used to train and
to the specific job market being analyzed. For example, if test the system is critical to its overall effectiveness. In particular,
the system is being developed for a software development having a diverse and representative sample of resumes can help
ensure that the system can accurately handle a range of different
job titles, skill sets, and other variables.
4.1.4 The impact of machine learning algorithms: 4. Continuous learning: Future research can explore
Another key finding may be the impact of machine learning developing systems that continuously learn from user
algorithms on the overall effectiveness of the system. For feedback and improve over time.
example, using an algorithm that can learn from past parsing
successes and failures, or that can adapt to new job postings and 5. Conclusion:
resume formats, may result in a more accurate and effective The research on resume parsing for software job profiles has
system overall. demonstrated the potential benefits of using automated systems
for resume screening and ranking. The study identified several
4.2 Implications of Recruitment Practices limitations and future directions for research, including the need
the findings suggest that the use of a resume parsing system can for more comprehensive data sets and the potential for bias in the
help to reduce unconscious bias in the recruitment process, as the design and implementation of resume parsing systems. Despite
system can be programmed to prioritize objective criteria such as these limitations, the implications of the research suggest that
skills and experience over subjective factors such as education resume parsing systems have the potential to revolutionize the
level or gender. This can lead to a more diverse and inclusive recruitment process for software job profiles. By improving
candidate pool and ultimately, a more effective recruitment efficiency, reducing bias, and improving the overall quality of
process. candidate selection, these systems can play an important role in
the hiring process for software jobs. Overall, the research
The implications of the research on resume parsing for software highlights the importance of continued development and
job profiles can be significant for recruitment processes. The improvement of resume parsing systems for software job profiles,
findings suggest that the use of automated systems for resume and the potential benefits that these systems can bring to the
screening and ranking can save time and resources for recruiters, recruitment process.
while also improving the quality of candidate selection. By using
a resume parsing system, recruiters can ensure that they are able
to quickly and efficiently identify candidates with the necessary 6. REFERENCES
skills and experience for the job.

4.3 Limitations and future scope [1]. F. Ciravegna, “Adaptive information extraction from text by
rule induction and generalisation,” in Proceedings of the 17th
4.3.1 Limitations:
International Joint Conference on Artificial Intelligence
1. Quality of data: One of the main limitations of such
(IJCAI2001), 2001.
systems is the quality of data used. The accuracy of the
parsing system depends on the quality of the resume and
job description dataset. Incomplete, outdated or [2]. A. Chandel, P. Nages h, and S. Sarawagi, “Efficient batch
irrelevant data can impact the accuracy of the system.
top-k search for dictionary-based entity recognition,” in
2. Bias: Another limitation is the possibility of bias in the
Proceedings of the 22nd IEEE International Conference on Data
system due to the input data. For example, if the training
data includes a disproportionate number of resumes or Engineering (ICDE), 2006.
job descriptions from certain regions or industries, the
system may be biased towards those regions or
industries. [3]. S. Chakrabarti, Mining the Web: Discovering Knowledge
from Hypertext Data. Morgan-Kauffman, 2002
3. Ambiguity in Job Titles and Skills: The ambiguity of
job titles and skills is also a limitation for such systems.
In software job profiles, new job titles are introduced [4]. M. J. Cafarella, D. Downey, S. Soderland, and O. Etzioni,
regularly, and it can be difficult for parsing systems to
“KnowItNow: Fast, scalable information extraction from the
accurately identify the exact job responsibilities and
required skills. web,” in Conference on Human Language Technologies
(HLT/EMNLP), 2005.
4.3.2 Future Scope:
1. Multimodal input: One direction for future research is to
include multimodal input like images and video along [5]. M. J. Cafarella and O. Etzioni, “A search engine for natural
with the traditional text input. This will provide a richer language applications,” in WWW, pp. 442–452, 2005.
dataset and can help in building more accurate models.
2. Enhanced machine learning models: With advancements [6]. https://www.ijircce.com/upload/2016 /april/218_ Intellig
in machine learning, future research can explore using
more advanced algorithms like deep learning and neural ent.pdf
networks to improve the accuracy of parsing systems.
3. Contextual analysis: Another direction is to incorporate [7]. https://www.tutorialspoint.com/compiler_design/images
contextual analysis by using natural language /token_passing.jpg
processing and sentiment analysis to extract additional
information from resumes and job descriptions.
[8]. http://www.nltk.org/book/tree_images/ch08-tree-6.png