e-ISSN: 2582-5208

International Research Journal of Modernization in Engineering Technology and Science

( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:04/Issue:07/July-2022 Impact Factor- 6.752 www.irjmets.com
MACHINE LEARNING TECHNIQUES FOR STRESS PREDICTION IN
WORKING EMPLOYEES SCIENCE & ENGINEERING
Altamash A Maniyar*1, Jeevan Kumar SH*2, Nithej N*3,
Ramya HK*4, Aishwarya T*5
*1,2,3,4,5Computer Science & Engineering, Vidyavardhaka College Of Engineering,
Mysuru, Karnataka, India.
ABSTRACT
The modern world is filled with stress. Everyone is under pressure in a situation for one of two reasons. A
person's pressure is affected by a variety of factors. Representatives in IT are more likely to be under pressure
due to work pressure, overburdening, higher worker mastery, and so on. When a person is stressed, it can lead
to a variety of mental health issues such as depression, anxiety, somatization, lack of concentration, and so on. It
can sometimes be fatal. As a result, it is necessary to identify human stress at an early stage in order to provide
appropriate solutions and alleviate stress. There has been a lot of research done on stress prediction. Many
research papers use Machine Learning techniques to predict stress, and many papers use IOT-based sensors to
extract the features needed for stress prediction. Many papers simply present the concept of stress prediction
without any implementation. There are some research papers that include implementation. These
implementation papers make use of ready-made tools such as the WEKA tool, the R tool, the Rapid Miner, or
programming languages such as PYTHON or R. It is simple to predict stress using these ready tools and
languages because they support ready libraries for stress prediction. Data science techniques are effective at
processing training datasets and can predict human stress in less time and with higher accuracy.
Keywords: Data Science, Machine Learning, Stress, Information Technology Profession.
I. INTRODUCTION
The modern world is flooded with IT, and IT companies are being greeted with new extensions and requests.
Representatives are bound to face pressure as a result of the changing way of life and working societies.
Frameworks identify components that have a significant impact on anxiety feelings. Stress was identified in
relation to orientation, family ancestry, and the availability of medical benefits in the workplace. Recognizing
the pressure on representatives allows us to devise a few strategies for dealing with it and creating a much
more pleasant work environment for their representatives.
Several research works make use of a variety of constraints, such as preferences, age, family history, provided
medical benefits, shared information about illness, technical institution, technical job, acquiring holidays, and
so on. Artificial Intelligence calculations are used in research to determine an employee's stress level. The
primary goal of each of these research articles is to identify the gambling factors that influence the worker's
emotional wellness.
II. PROPOSED SYSTEM
1. The presented system is designed for commercial use.
2. Systems identify variables that have a serious influence on stress levels.
3. Stress was discovered based on sex, family background, and the existence of medical advantages in the
professional sector.
4. Understanding employee stress allows us to build policies or treatments to reduce stress levels and focus on
improving the work atmosphere.
5. Gender, age, economic difficulties, family conflicts, work schedule, learning methods, medical issues,
conflicts with co-workers, pressure, regularity, and communication are all major parameters considered by the
system.
6. The system uses Machine learning techniques to detect stress levels.

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[2958]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:04/Issue:07/July-2022 Impact Factor- 6.752 www.irjmets.com
7. The project can be implemented as a real-time business application. We chose Visual Studio and SQL Server
for application development because they are more real-time application supportive.
8. The project's primary aim is to discover risk factors that affect on the mental state of employees that work.
9. The project also makes practical suggestions to the working staff based entirely on the stress levels.
10. The project's goal is to identify causative factors and forecast future levels of employee emotional turmoil.
We devise methods to reduce stress levels so that working employees' performance can be improved.

Fig 1. System Architecture of the Proposed System

III. METHODOLOGY
Step 1: Data Collection (Stress Data)
During this stage of the employee stress prediction process, we collect stress data. Gender, age, financial, family,
working hours, learning method, physical well-being, favouritism fix, co-worker issues, pressure, regular
communication, and so on are all parameters that have been collected from a variety of sources.
Step 2: Data Preparation
Following the evaluation of the stress data, only important data was collected. Data required for computation is
utilized and fragmented according to parameters. Because complete data is not required for processing and
entering all data would take too long, data extraction is performed.
Step 3: Specify Constraints
The stress factors used to evaluate stress levels are retrieved. Gender, age, financial issues, family issues, work
schedules, learning methods, medical issues, favouritism issues, co-worker issues, pressure issues, routine
issues, communication issues, and so on.
Step 4: ML Algorithms
 Supervised Learning
Supervised learning is a machine learning method that is based on training data sets that contain predicted
responses.
 The KNN Algorithm And Naive Bayes, As Well As The Decision Tree
1. A Strong Classifier.
2. It Performs Well With Fewer Parameters As Well As Many Parameters.
3. Effective With Both Small And Big Data Sets.
4. More Precise Outcomes.
Step 5: Stress Prediction.
Based on the factors, the system employs a machine learning algorithm to predict the stress levels of employees
that work. We are using the KNN algorithm and the Naive Bayes algorithm to identify stress levels.

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[2959]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:04/Issue:07/July-2022 Impact Factor- 6.752 www.irjmets.com
Step 6: Obtaining Results
To determine the correctness of the algorithm, we divide the training datasets into training and testing
datasets. 90% of the records in the datasets are considered training datasets, while 10% are considered testing
datasets.
Step 7: Visual Representation
The output is displayed to the user via the GUI.
Users:
 Administrator - the person in charge of the entire application. The one with complete authority. Admin can
respond to queries submitted by the employees.
 Employees - Working employees are the ones who receive services and can submit questions to
administrator.
Admin Modules
1. Admin Login Module - By entering the admin id and password, admin gains access to the application.
2. Add Employees - Admin will add all employees from various departments here.
3. Set Id and Password - Here, the administrator can create a distinct id and passcode for each person
employed.
4. View Employees - Here, the administrator can see all of the employees in the organization, as well as their
personal information.
5. Manage Dataset - Administrator can view the training datasets used in the project, which was imported
from an excel file, in this section.
6. Stress Prediction Module - This is the main module where the platform identifies the stress issues that
working employees experience.
 In this case, we use a data science technique known as "supervised learning" KNN algorithm. (YES OR NO
PREDICTION)
 In order to analyze and evaluate the system's accuracy and performance, administrator can predict output
for the testing datasets stored in an excel sheet.
 Result evaluation (correctness and effectiveness) for the testing datasets extracted from an excel sheet.
7. Solution Module – Administrator can recommend and manage solutions for each employee in the
organization based on their stress level.
 All countermeasures for every stress level unit can be managed by the administrator.
8. Queries - Administrators can view employee queries and respond. The Queries module also includes
pending and answered queries modules for viewing both types of queries.

Fig 2. Admin module dataflow diagram

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science
[2960]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:04/Issue:07/July-2022 Impact Factor- 6.752 www.irjmets.com
Employee Modules
1. Employee Login Module - Users can access the system by entering their unique employee id and password.
2. Employee Stress Prediction Module – The user will enter all of the necessary parameters for stress level
prediction.
3. Stress Prediction Module - This is the main module where the system identifies the stress issues that
working employees face. In this case, we use a data science technique known as "supervised learning" KNN
algorithm.
4. Solution Module - The system will display remedies defined by the administrator for working employees'
stress issues.
5. Queries Module – It is same as in the administrator query module, the only difference is the Post queries
module.
 Post Queries – If any questions are submitted to the administrator, the user can view the responses.

Fig 3. Employee module dataflow diagram

Project Outcome
Input: For computation, the system makes use of many parameters such as gender, age, family background,
provided medical benefits, sharing about diseases, tech company, tech role, obtaining leave, and so on, as well
as old data-sets.
Output: Employees are labelled as Stressed or Stress-Free dependent on their stress levels.

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[2961]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:04/Issue:07/July-2022 Impact Factor- 6.752 www.irjmets.com

Fig 4. Proposed System flow chart

IV. CONCLUSION
Mental stress is a serious problem in today's society. It has been concluded that different techniques are used to
detect stress. There are also several datasets available for identifying stress. Gender, family medical history of
illness, and whether or not an organization provides mental health benefits to their employees were discovered
to be more important than other factors in determining whether or not a person can develop mental health
issues. People who work in a tech company, even if their role is not technically related, are slightly more likely
to develop stress, according to our findings. These insights can be used by corporations to develop better HR
policies for their employees. Moreover, ensemble methods such as boosting and KNN produced the highest
classification precision and accuracy. With an accuracy of 87%, the use of Machine Learning techniques for
identifying stress and mental health disorders yields significant results that can be investigated further, thereby
meeting the project's goal.
V. REFERENCE
[1] Bhattacharyya, R., & Basu, S. (2018). India Inc looks to deal with rising stress in employees. Retrieved
from ‘The Economic Times’
[2] OSMI Mental Health in Tech Survey Dataset, 2017 from Kaggle
[3] Van den Broeck, J., Cunningham, S. A., Eeckels, R., & Herbst, K. (2005). Data cleaning: detecting,
diagnosing, and editing data abnormalities. PLoS medicine, 2(10), e267.

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[2962]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:04/Issue:07/July-2022 Impact Factor- 6.752 www.irjmets.com
[4] Shwetha, S, Sahil, A, Anant Kumar J, (2017) Predictive analysis using classification techniques in
healthcare domain, International Journal of Linguistics & Computing Research, ISSN: 2456-8848, Vol. I,
Issue. I, June-2017
[5] Tomar, D., & Agarwal, S. (2013). A survey on Data Mining approaches for healthcare. International
Journal of Bioscience and Biotechnology, 5(5), 241-266.
[6] Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., &Vanderplas, J. (2011).
Scikit-learn: Machine learning in Python. Journal of machine learning research, 12(Oct), 2825-2830.
[7] Gender and Stress. (n.d.). Retrieved from APA press release 2010
[8] Vidit Laijawala, Aadesh Aachaliya, Hardik Jatta, Vijaya Pinjarkar. Classification Algorithms based
Mental Health Prediction using Data Mining (2020).
[9] U Srinivasulu Reddy, Aditya Vivek Thota, A Dharun. Machine Learning Techniques for Stress Prediction
in Working Employees (2020).
[10] Monisha S, Meera R, Vijay Swaminath.R, Dr. Arun Raj L. Predictive Analysis of Student Stress Level
Using Naive Bayesian Classification Algorithm (2020).
[11] Fang Li. Research on the College Students Psychological Health Management based on Data Mining and
Cloud Platform (2016).
[12] Jigang Zheng, Jingmei Zhang. The Application of Association Rules Mining in the Analysis of Students
Test Scores (2016).
[13] Jenny K. Hyun, Brian C. Quinn, Temina Madon, Steve Lustig. Graduate Student Mental Health: Needs
Assessment and Utilization of Counseling Services (2014).
[14] Shivendra Jena, Harish Chandra Tiwari. Stress and mental health problems in 1st year medical
students: a survey of two medical colleges in Kanpur, India (2014).
[15] Honglei Zhu, Zhigang Xu. An Effective Algorithm for Mining Positive and Negative Association Rules
(2008).

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[2963]