Employee Future Prediction
Employee Future Prediction
Volume 6, Issue 11, November – 2021 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Employee Future Prediction
Sai Chandan P Reddy
Priyanshi Thakur
Electronics and instrumentation engineering
Ms Ramaih Institute of technology
Bangalore, India
Abstract:- The biggest asset that a company has is not his increasingly efficient worker attrition classifiers. the use of
business or the place or any other thing, but it’s the greater numerous datasets or certainly to replace it periodically,
employee. The employees pave the path for the success of the software of feature [3] engineering to identify new sizable
the company and all the administration that is needed to run traits from the dataset and the supply of additional statistics on
the company. But due to many reasons employee are not personnel would improve the overall information of the motives
happy with the work or the company they work in and this why personnel leave their agencies and, consequently, increase
tends or results for them to leave the company or finding a the time available to personnel departments to assess and plan
new company. It is one of the most important difficulties the duties required to mitigate this risk (e.g., retention activities,
that company owners confront when their organizations employee substitution and/or venture redistribution).[4]
lose their most brilliant personnel. A good employee is predicts the accuracy of five base models and then combines
always a valuable asset to the company, and their departure them to create a more powerful predictor model, which is
can result in a variety of issues, including financial losses, known as ensemble learning. The research uses a combination
poor overall performance, and the loss of accumulated of decision trees and linear regression to achieve an accuracy of
expertise. Furthermore, compared to recruiting current 86.39 percent, outperforming Adaboost and Random Forest, as
employees, employing new personnel is significantly more well as SVM.
expensive, stressful, and time-consuming. Hence, the
authors propose a system to predict the future of an III. METHODOLOGY
employee in a company taking in consideration various
factors and using the algorithm to get the prediction with a A. Data Preparation
great accuracy. The information was taken from the well-known employee
future prediction dataset on the internet. Due to the scarcity of
Keywords:- Prediction Algorithm, Machine Learning, data, the only method to run a model and get a forecast was to
Employee Future Prediction, Gradient Boosting, Attrition gather data from a reliable source. The dataset contains a variety
System. of properties like education, gender, age, city, payment tier,
joining year, leave or not, ever benched etc which can be seen
I. INTRODUCTION in figure 1. The dataset is quite big and contains ~4650 entries
which is taken from the employees in a company and 9 features
In many of the company’s employee, future has become a along with it.
major. It’s the responsibility of the HR (Human Resource)
Manager to fulfil necessary needs of their employs. Employs
generally leave the companies and to fulfil that particular
position they have to assign another person for that particular
position, which is not always possible for them. Since the
worker's quality and skills represents the companies the growth
and competitive advantage. [1] And there has been a method
developed called Predicting Employee Attrition. The purpose of
this is to support decisions are not based on personal aspects but
on unbiased data analysis. The goal of this work is to analyse
how purpose factors influence employee attrition, in order to
identify the man causes that contribute to a worker’s decision to
leave a company, and to predict whether a particular employee
will leave the company. Fig 1: Features/Attributes Overview
II. LITERATURE OVERVIEW B. Data Preprocessing
Preprocessing is a great approach to improve data because
There few steps that are being used for performing ML both the actual world and our data have flaws. The speed of the
processing, and they are Data Preprocessing, Data Analysis, procedure is determined on whether the data has been
Model Training, Model Validation, Model Prediction, preprocessed. The better the model that is utilized, the more
Visualization of results [2] The consequences received from the preprocessing that is done. Before eliminating the ID column,
data analysis constitute a starting point in the improvement of the author checks for all null values, which will have no effect
on the results. After removing the null values, the author
IJISRT21NOV481 www.ijisrt.com 201
Volume 6, Issue 11, November – 2021 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
performed a describe function to get to know the dataset more,
which would also be beneficial in doing the visualization as it
would give the right relation to perform. The figure 2 shows the
describe table of the dataset.
Fig 2: Describing the dataset Fig 5: Education Visualisation
C. Feature Visualisation
The algorithm's features are crucial for producing reliable
results. Visualization helps us to see all the different factors and
how they influence the results. The figure 3 shows a heat map
between the features/attributes and which signifies Our Target
variable ('LeaveOrNot') has a very modest negative connection
with PaymentTier and Age.The target variable ('LeaveOrNot')
exhibits a very minor positive connection with 'JoiningYear.'
Figure 4 shows that Most of the Employees are having payment
Tier 3 By observing the trend it seems that 'PaymentTier'
Catergory is an Ordinal Variable Where, Tier 3 > Tier 2 > Tier
1. Figure 5 shows that most of the employees have a bachelors
degree but the employees with masters degree have left more.
And in figure 6 it shows that young people have left early due
Fig 6: Age GroupVisualisation
to more job opportunities out in the market.
D. Model Architecture
Gradient boosting is similar to random forest in that it
groups weak learners together to build a stronger one. However,
gradient boosting differs from random forest in that it adds
predictors in a sequential manner, preventing additional
improvement. The gradient booster uses the gradient of descent
to help discover and fix inaccuracies in learners' predictions. In
compared to decision trees and linear regression, it is memory
efficient and fast, but it fails to visualise perfectly. Gradient
boosting framework is used which is used to handle large size
data which in this case is a dataset of ~4650 and takes very low
memory to run. It also provides a support for the GPU learning
and thus is widely use by the data scientists and hence helps our
title to increase the accuracy.
IV. EXPERIMENTAL RESULTS
Fig 3: Heat map of Attributes
The procedure followed from data preparation to
preprocessing has helped the dataset to get better for the model
and has played a vital role in the accuracy, and the algorithm
used in the method provided us with an accuracy of ~ 93 percent
which is a very improved accuracy then the work that has been
performed in the earlier researches done by different people.
V. CONCLUSION
Employees are the main asset to the company and the
company can’t run without them and they play an important role
in shaping the company and sending the company to next
Fig 4: Payment Tier Visualisation heights i.e making it successful hence the algorithm proposed
IJISRT21NOV481 www.ijisrt.com 202
Volume 6, Issue 11, November – 2021 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
by the authors help the HR managers to know what should be [12]. . C. Patro, S. A. Zaidi, A. Dixit and M. Dixit, ”A Novel
improved if an employee is thinking of quitting or finding a new Approach to Improve Employee Retention Using Machine
job and by this they would not loose a credible employee to any Learning,” 2021 10th IEEE International Conference on
other rival company, So the accuracy given by the algorithm Communication Systems and Network Technologies
would make an impact on the present systems and increase the (CSNT), 2021, pp. 680-684, doi:
predictions by a very high percentage, 10.1109/CSNT51715.2021.95096013.
[13]. R. S. Shankar, J. Rajanikanth, V. V. Sivaramaraju and K.
REFERENCES V. S. S. R. Murthy, ”PREDICTION OF EMPLOYEE
ATTRITION USING DATAMINING,” 2018 ieee
[1]. P. Ajit, "Prediction of Employee Turnover in international conference on system, computation,
Organizations using Machine Learning Algorithms". automation and networking (icscan), 2018, pp. 1-8, doi:
[2]. M. Guirao, "Predicting Employee Attrition with Machine 10.1109/ICSCAN.2018.8541242
Learning". [14]. V. Shah and S. Modi, "Comparative Analysis of
[3]. F. Fallucchi, E. W. D. Luca and R. Giuliano, "Predicting Psychometric Prediction System," 2021 Smart
Employee Attribution Using Machine Learning".. Technologies, Communication and Robotics (STCR),
[4]. Qutub, Aseel, et al. ”Prediction of Employee Attrition 2021, pp. 1-5, doi: 10.1109/STCR51658.2021.9588950.
Using Machine Learning and Ensemble Methods.” Int. J.
Mach. Learn. Comput 11 (2021)
[5]. H. Mallafi and D. H. Widyantoro, "Prediction modelling
in career management," 2016 International Conference on
Computational Intelligence and Cybernetics, 2016, pp. 17-
21, doi: 10.1109/CyberneticsCom.2016.7892560.
[6]. T. T. Weeramanthrie, C. N. Thilakumara, K. N. A. C.
Wijesiri, N. I. Fernando, S. Thelijjagoda and A. Gamage,
"ARROW: A web-based employee turnover analysis tool
for effective human resource management in large-scale
organizations," 2017 National Information Technology
Conference (NITC), 2017, pp. 136-140, doi:
10.1109/NITC.2017.8285665.pp. 178392-178400, 2019,
doi: 10.1109/ACCESS.2019.2956859.
[7]. Wang Hong and Sun Yanshen, "Predicting employee
relations in enterprises by bi-directional link mining,"
Proceeding of the 11th World Congress on Intelligent
Control and Automation, 2014, pp. 3729-3734, doi:
10.1109/WCICA.2014.7053337.s (ICIIECS), 2017, pp. 1-
3, doi: 10.1109/ICIIECS.2017.8276033.
[8]. R. Chakraborty, K. Mridha, R. N. Shaw and A. Ghosh,
"Study and Prediction Analysis of the Employee Turnover
using Machine Learning Approaches," 2021 IEEE 4th
International Conference on Computing, Power and
Communication Technologies (GUCON), 2021, pp. 1-6,
doi: 10.1109/GUCON50781.2021.9573759.
[9]. J. Liu, T. Wang, J. Li, J. Huang, F. Yao and R. He, "A
Data-driven Analysis of Employee Promotion: The Role
of the Position of Organization," 2019 IEEE International
Conference on Systems, Man and Cybernetics (SMC),
2019, pp. 4056-4062, doi: 10.1109/SMC.2019.8914449.
[10]. Jain, P.K., Jain, M. Pamula, R. Explaining and predicting
employees’ attrition: a machine learning approach. SN
Appl. Sci. 2, 757 (2020). https://doi.org/10.1007/s42452-
020-2519-4
[11]. S. Modi and M. H. Bohara, ”Facial Emotion Recognition
using Convolution Neural Network,” 2021 5th
International Conference on Intelligent Computing and
Control Systems (ICICCS), 2021, pp. 1339-1344, doi:
10.1109/ICICCS51141.2021.9432156
