|| Volume 6 || Issue 5 || May 2021 || ISSN (Online) 2456-0774

INTERNATIONAL JOURNAL OF ADVANCE SCIENTIFIC RESEARCH

AND ENGINEERING TRENDS

SALARY PREDICTION USING MACHINE LEARNING

Prof. D. M. Lothe1, Prakash Tiwari2, Nikhil Patil3, Sanjana Patil4, Vishwajeet Patil5
Professor, Information Technology, JSPM’s Rajarshi Shahu College of Engineering, Pune, India 1
UG Student, Information Technology, JSPM’s Rajarshi Shahu College of Engineering, Pune, India2345
dml.rscoe@gmail.com1, prakashtiwari468@gmail.com2, np0129999@gmail.com3, patilsanjanam@gmail.com4,
svishpatil@gmail.com
------------------------------------------------------ ***--------------------------------------------------
Abstract: - Machine learning is a technology which allows a software program to became more accurate at pretending more
accurate results without being explicitly programmed and also ML algorithms uses historic data to predicts the new outputs.
Because of this ML gets a distinguish attention. Now a day’s prediction engine has become so popular that they are
generating accurate and affordable predictions just like a human, and being using industry to solve many of the problems.
Predicting justified salary for employee is always being a challenging job for an employer. In this paper and proposing a
salary prediction model with suitable algorithm using key features required to predict the salary of employee.
Keywords: - Machine Learning, Linear regression, Model selection, Supervised Learning
---------------------------------------------------------------------***---------------------------------------------------------------------

I INTRODUCTION regression problem is when the output value is real or a

continuous value like salary.
Now days, Major reason an employee switches the company is
the salary of the employee. II LITERATURE REVIEW
Employees keep switching the company to get the expected 1) Susmita Ray," A Quick Review of Machine Learning
salary. And it leads to loss of the company and to overcome this Algorithms," 2019 International Conference on Machine
loss we came with an idea what if the employee gets the Learning, Big Data, Cloud and Parallel Computing (Com-IT-
desired/expected salary from the Company or Organization. In Con), India, 14th -16th Feb 2019 a brief review of various
machine learning algorithms which are most frequently used
this Competitive world everyone has a higher expectation and
to solve classification, regression and clustering problems.
goals. But we cannot randomly provide everyone their expected The advantages, disadvantages of these algorithms have been
salary there should be a system which should measure the discussed along with comparison of different algorithms
ability of the Employee for the Expected salary. (wherever possible) in terms of performance, learning rate
We cannot decide the exact salary but we can predict it by etc. Along with that, examples of practical applications of
using certain data sets. these algorithms have been discussed.[1]
A prediction is an assumption about a future event. 2) Sananda Dutta, Airiddha Halder, Kousik Dasgupta,”
Design of a novel Prediction Engine for predicting suitable
In this paper the main aim is predicting salary and
salary for a job” 2018 Fourth International Conference on
making a suitable user-friendly graph. So that an Employee can Research in Computational Intelligence and Communication
get the desired salary on the basis of his qualification and hard Networks (ICRCICN) - focused on the problem of predicting
work. For developing this system, we are using a Linear salary for job advertisements in which salary are not
regression algorithm of supervised learning in machine mentioned and also tried to help fresher to predict possible
learning. salary for different companies in different locations. The
corner stone of this study is a dataset provided by ADZUNA.
Supervised learning is basically a learning task of a model is well capable to predict precise value.[2]
learning function that maps an input to an output of given
3) Pornthep Khongchai, Pokpong Songmuang, “Improving
example. In supervised learning each example is pair having Students’ Motivation to Study using Salary Prediction
input parameter and the desired output value. System” - proposed prediction model using Decision tree
Linear regression algorithm in machine learning is a technique with seven features. Moreover, the result of the
supervised learning technique to approximate the mapping system is not only a predicted salary, but also the 3-highest
salary of the graduated students which share common
function to get the best predictions. The main goal of regression attributes to the users. To test the system’s efficiency, they set
is the construction of an efficient model to predict the up an experiment by using 13,541 records of actual graduated
dependent attribute from a bunch of attribute variables. A student data. The total result in accuracy is 41.39%.[3]

IMPACT FACTOR 6.228 WWW.IJASRET.COM DOI : 10.51319/2456-0774.2021.5.0047 199

 || Volume 6 || Issue 5 || May 2021 || ISSN (Online) 2456-0774
INTERNATIONAL JOURNAL OF ADVANCE SCIENTIFIC RESEARCH
AND ENGINEERING TRENDS
4) Phuwadol Viroonluecha, Thongchai Kaewkiriya,” Salary
Predictor System for Thailand Labour Workforce using Deep
Learning” - used Deep learning techniques to construct a model
which predicts the monthly salary of job seekers in Thailand
solving a regression problem which is a numerical outcome is
effective. We used five-month personal profile data from well-
known job search website for the analysis. As a result, Deep
learning model has strong performance whether accuracy or
process time by RMSE 0.774 x 104 and only 17 seconds for
runtime.[4]
III INTRODUCTION
In order to gain useful insights into the job recruitment, we
compare different strategies and machine learning models. The
methodology different phases like: Data collection, Data
cleaning, Manual feature engineering, Data set description,
Automatic feature selection, Model selection, Model training
and validation, Model comparison.
We are focusing to develop a system that will predict the salary
based on different parameters used in company and above-
mentioned methodology phases. Some of the parameters we
collected from company data are: Job Type: CFO, CEO, Senior,
vice president, manager

1. Degree: Doctoral, Bachelors, Masters, High School Figure -1: System architecture diagram
2. Major: Math, Literature, Engineering, Business, Step 3:
Physics, Chemistry 1. Visualize Linear regression results
3. YearsExperience:
4. Industry: Health, Service, Finance, Product, Web,
Education
Miles from Metropolis:
5. Salary:
The calculations that will be performed for working of this
proposed system to predict the salary with results:
Step 1: In step 1 we consider only Years’ Experience vs
Salary to create a base Model.
Here X is the independent variable which is the “years’
Experience”. And y is the dependent variable which is the
“Salary”.
Step 2:

1. Fit linear regression model to database Chart -1: Base model result

2. Firstly, building a simple Linear Regression model to 2. So, we see here that the line created by our model is
quite accurate.
see what prediction it makes. 3. The accuracy of our model is now around 96% to
3. We will be using the LinearRegression class from the 98%.
library sklearn.linear_model. We create an object of To create a basic training model, two variables will be assigned
the LinearRegression class and call the fit method for the model to use. Twenty percent of the training data will be
split into testing data that we can use to test the model with data
passing the X and y. for which the salaries are already known.

IMPACT FACTOR 6.228 WWW.IJASRET.COM DOI : 10.51319/2456-0774.2021.5.0047 200

 || Volume 6 || Issue 5 || May 2021 || ISSN (Online) 2456-0774
INTERNATIONAL JOURNAL OF ADVANCE SCIENTIFIC RESEARCH
AND ENGINEERING TRENDS
Step 4:
1. Now our main challenge is to add more parameters
and maintain the accuracy
2. Next, we visualize each categorical (jobType, degree,
major, industry, yearsExperience,
milesFromMetroplis) feature to see which features
could be good predictors of salary.
3. So, by visualizing each category we come to know
that, yearsExperience yearsExperience has the highest
correlation with salary. jobType also seems to be
correlated with salary.

Step 5:
1. Create baseline Model
2. Baseline model is created on the dataset that contains
Figure -3: Polynomial transformation result
all features using Linear Regression
Where we will 80% data for training and 20% data to Here model updated using polynomial transformation.
check our model. Step 7: Predict the MSE and accuracy of new model
3. Mean squared error (MSE) will be evaluated now
along with accuracy to evaluate the baseline model's
performance.

Figure -4: MSE after applying Polynomial Transformation

2. Visualize the data
Figure -2: Base model MSE result
After this stage the MSE is very high. It is 384. Now our aim is
to reduce it.
So, to reduce MSE to less than 360. We can use:
a. Apply Polynomial Transformation
b. Use Ridge Regression
c. Use Random Forest
From these 3 ways we are continuing with polynomial
transformation which actually reduced the MSE.
Step 6:
1. We will be using the Polynomial Features class from
the sklearn.preprocessing library for this purpose.
When we create an object of this class — we have to
pass the degree parameter.
2. Let’s begin by choosing degree as 2. Appy Polynomial
Transformation.
If we have 2 labels named x1 and x2, then after
applying polynomial transformation of degree 2 the
new features will be:
1, x11, x21, x12, x22, (x1 * x2)
3. Fit and Transform the variables with 2nd order
polynomial and then create a Linear Regression model Chart -2: Baseline model result after adding Polynomial
on the new data. Transformation

IMPACT FACTOR 6.228 WWW.IJASRET.COM DOI : 10.51319/2456-0774.2021.5.0047 201

 || Volume 6 || Issue 5 || May 2021 || ISSN (Online) 2456-0774
INTERNATIONAL JOURNAL OF ADVANCE SCIENTIFIC RESEARCH
AND ENGINEERING TRENDS
3. The new MSE is 357.
4. Linear regression with second order polynomial
transformation gave best predictions with MSE of 357
and accuracy of 76%. This meets the goal of reducing
MSE to below 360.

IV CONCLUSIONS
In this paper we proposed a salary prediction system
by using a linear regression algorithm with second order
polynomial transformation. For the proper salary prediction, we
found out most relevant 5 features. The result of the system is
calculated by suitable algorithm by comparing it with another
algorithms in terms of standard scores and curves like the
classification accuracy, theF1score, the ROC curve, the
Precision-Recallcurve etc. We compared algorithms only for
the basic model which only two attributes. Moreover, we
continued with basic model and found out the most appropriate
method to add more attribute and with highest accuracy of
76%.
In future work, we would like add graphical user
interface to system and try to save and reuse trained model.
ACKNOWLEDGEMENT
We would like to express our sincere gratitude to several
individuals and organizations for supporting me throughout my
Project. First, I wish to express my sincere gratitude to my
supervisor, Professor D. M. Lothe, for his enthusiasm, patience,
insightful comments, helpful information, practical advice and
unceasing ideas that have helped us tremendously at all times in
our research. Without her support and guidance, this project
would not have been possible. I could not have imagined
having a better supervisor in our study.

REFERENCES
[1] Susmita Ray," A Quick Review of Machine Learning
Algorithms," 2019 International Conference on Machine
Learning, Big Data, Cloud and Parallel Computing (Com-
IT-Con), India, 14th -16th Feb 2019
[2] Sananda Dutta, Airiddha Halder, Kousik Dasgupta,”
Design of a novel Prediction Engine for predicting suitable
salary for a job” 2018 Fourth International Conference on
Research in Computational Intelligence and
Communication Networks (ICRCICN).
[3] Pornthep Khongchai, Pokpong Songmuang, “Improving
Students’ Motivation to Study using Salary Prediction
System” 2016 13th International Joint Conference on
Computer Science and Software Engineering (JCSSE)
[4] Phuwadol Viroonluecha, Thongchai Kaewkiriya,” Salary
Predictor System for Thailand Labour Workforce using
Deep Learning” The 18th International Symposium on
Communications and Information Technologies (ISCIT
2018)
[5] Mangui Wu, Shunmin Shu,” Top Management Salary,
Stock Ratio and Firm Performance: A Comparative Study
of State-owned and Private Listed Companies in China”

IMPACT FACTOR 6.228 WWW.IJASRET.COM DOI : 10.51319/2456-0774.2021.5.0047 202