A Project

ON

SALARY PREDICTION USING REGRESSION

TECHNIQUES

Submitted in partial fulfillment of the

requirement for the award of the degree of

B.Tech Computer Science Engineering

Name of Supervisor: Abhay Kumar

Assistant Professor

Submitted By:

Name of Student: PULENDRA KUMAR YADAV

Admission Number: 19SCSE1010413
Application Number: 19021011593
Sec 09

&

Name of Student: RIKESH KUMAR

Admission Number: 19SCSE1010385
Application Number: 19021011565
Sec 09

SCHOOL OF COMPUTING SCIENCE AND ENGINEERING DEPARTMENT OF

COMPUTER SCIENCE AND ENGINEERINGGALGOTIAS UNIVERSITY,
GREATER NOIDA
INDIA MONTH, YEAR

Electronic copy available at: https://ssrn.com/abstract=3990877

SALARY PREDICTION USING REGRESSION
TECHNIQUES

Abstract:
The goal of this paper is to predict salary of a person after a
certain year. The graphical representation of predicting salary
is a process that aims for developing computerized system to
maintain all the daily work of salary growth graph in any field
and can predict salary after a certain time period. This
application can take the database for the salary system from the
organisation and makes a graph through this information from
the database. It will check the salary fields then import a graph
which helps to observe the graphical representation. And then
it can predict a certain time period salary through the
prediction algorithm. It can also be applied in some other
effective prediction also.

Electronic copy available at: https://ssrn.com/abstract=3990877

 Title Page
No.
Abstract I
List of Table II
List of Figures III
Chapter 1 Introduction
1.1 Introduction

Chapter 2 Literature Survey/Project Design

2.1 METHODLOGY
2.2 EXPERIMENTATION
2.3 RESULT AND DISCUSSION
2.4 CONCLUSION

Electronic copy available at: https://ssrn.com/abstract=3990877

 I. INTRODUCTION  User can give any point to get the salary
through the program.
A prediction is an assumption about a future event.  Salary of the employees can be observed to
A prediction is sometimes, though not always, is give them a particular field according to their
based upon knowledge or experience. Future qualifications.
events are not necessarily certain, thus confirmed
exact data about the future is in many cases are The graphs through the Linear and polynomial graphs
impossible, a prediction may be useful to help in are displayed to detect the salary and position levels.
preparing plans about probable developments. In
this paper salary of an employee of an organization II. METHODOLOGY
is to be predicted on basis of previous salary
growth rate. Here history of salary has been Machine Learning (ML) [1] is basically that field of
observed and then on basis of that salary of a computer science with the help of which computer
person after a certain period of time it can be systems can provide sense to data in much the same
calculated automatically. way as human beings do. In simple words, ML [2] [3]
is a type of artificial intelligence that extracts patterns
In this paper the main aim is predicting salary and out of raw data by using an algorithm or method. The
making a suitable user-friendly graph. From this key focus of ML is to allow computer systems to
prediction the salary of an employee can be learn from experience without being explicitly
observed according to a particular field according programmed or human intervention.
to their qualifications. It helps to see the growth of Linear Regression [4] is an algorithm of machine
any field. It can produce a person’s salary by learning based on supervised learning scheme.
clustering and predict the salary through the graph. Linear regression [6] carries out a task that may
Using linear regression and polynomial regression predict the value of a dependent variable (y) on
it makes a graph. This graph helps to predict the basis of an independent variable (x) that is given.
salary for any positions. Therefore, this kind of regression technique looks
for a linear type of relationship between input x
The application is aimed to develop to maintain a and output y.
day-by-day monitoring to see the graphical Polynomial Regression [5] [7] is a form of linear
medium of any field (salary or experiences as well regression in which the relationship between the
as designation, etc.). A polynomial term: a independent variable x and dependent variable y is
quadratic (squared) or cubic (cubed) term turns a modelled as an nth degree polynomial. Polynomial
linear regression model into a curve. But because regression [8] fits a nonlinear relationship between
it is the data X that is squared or cubed, not the the value of x and the corresponding conditional
Beta coefficient, it still qualifies as a linear model. mean of y, denoted E(y |x).
This makes it a nice and straightforward way to This application provides a Salary graph
model curves without having to model representation that is mainly done by polynomial
complicated nonlinear models .One common regression statistics, polynomial regression [8] is a
pattern within machine learning is to use linear form of regression analysis which represents the
models trained on nonlinear functions of the data. relationship between the independent variable x
This approach maintains the generally fast and the dependent variable y and that is modelled
performance of linear methods while allowing as the nth degree of polynomial in x. Polynomial
them to fit a much wider range of data. That helps regression is suitable for a nonlinear type of
for the curving design. relationship between the value of x and the
correlating conditional mean of y, represented as
E(y |x). Although polynomial regression fits a
It will help the employee as per following ways:
nonlinear model to the data, as a statistical
estimation problem it is linear, in the sense that the
 Helping to see the growth at any field.
regression function E(y | x) is linear in the
 With the help of machine learning it can easily
unknown parameters that are estimated from the
produce a graph.
data. For this reason, polynomial regression is
 Marketing easy to estimate the salary between
x-y axis.

Electronic copy available at: https://ssrn.com/abstract=3990877

 assumed to be a special case of multiple linear
regressions.
Curve fitting is a method of building a curve, or to
represent a mathematical function that is optimally III. EXPERIMENTATION
suitable to a series of data points, and possibly it is This prediction has been implemented through the
subject to constraints. Curve fitting can involve following approaches:
either interpolation, where an exact relevant to the
data is needed, or smoothing, in which a "smooth" • numpy:- NumPy is a Python package
function is created that is approximately suitable which stands for 'Numerical Python'. It is
the data. A related topic is regression analysis, the core library for scientific computing,
which focuses more on questions of statistical which contains a powerful n-dimensional
inference such as how much uncertainty is present array object, provide tools for integrating
in a curve that is fit to data observed with random C, C++ etc. It is also useful in linear
errors. Fitted curves can be used as an aid for data algebra, random number cap ability etc.
visualization, to infer values of a function where
no data are available, and to summarize the • matplotlib.pyplot:- matplotlib.pyplot is a
relationships among two or more variables. collection of command style functions
Extrapolation refers to the use of a fitted curve that make matplotlib work like
beyond the range of the observed data, and is MATLAB. Each pyplot function makes
subject to a degree of uncertainty since it may some change
reflect the method used to construct the curve as to a figure: e.g., creates a figure, creates a
much as it reflects the observed data. Following is plotting area in a figure, plots some lines
the descriptions of the method that the work has in a plotting area, decorates the plot with
been done. labels, etc.
• pandas:- pandas is a kind of software
Proposed Method for Salary Prediction: library written for the Python
programming language for data
Step 1: Salary data have been taken from dataset. manipulation and analysis. In particular,
it offers data structures and operations for
Step 2: Then the points corresponding to the salary manipulating numerical tables and time
data of an individual person have been plotted in the series. It is free software released under
graph. The data are initialized in pandas (ascending, the three-clause BSD license.
descending, mixed-up). Taking the dataset from each
pandas field and from the pandas dataset we plotted the
• read_csv:- Python is a great language for
points on the graph as per number wise or input wise that
doing data analysis, primarily because of
came real dataset.
the fantastic ecosystem of datacentric
Step 3: After that we using linear regression for draw python packages. pandas is one of those
lines between the points. packages and makes importing and
analyzing data much easier.
Step 4: If the points are not in linear way then we
• Linear Regression :-
use polynomial regression for curving purpose. Linear
Through the clustering points we can make a smooth and Regression will be used to perform linear
curve path. and polynomial regression and make
predictions accordingly. Now, you have
Step 5: After then through the linear/polynomial graph two arrays: the input x and output y. You
should call .reshape() on x because this
through the x-y axis we can predict salary.
array is required to be two-dimensional,
or to be more precise, to have one column
Step 6: Also, we predict a person on future salary and as many rows as necessary,
position as per the graph goes. Only take a particular
person position, then the prediction answer be
• Polynomial features: - Generate a new
executed through the help of the graph. feature matrix consisting of all
polynomial combinations of the features

Electronic copy available at: https://ssrn.com/abstract=3990877

 with degree less than or equal to the Get data from companies,
specified degree. For example, if an input or Collect data from
sample is two dimensional and of the employees.
form [a, b], the degree-2 polynomial Survey operators assigned for salary strive to get the
features. most significant input data in every possible way.
There is no way to decide that which approach is
• Predict:-predict the values where find at correct. The first possibility may assure large
the time period that how much there in companies, where as the second choice is mainly for
that graph. comparative smaller companies.
First, a dataset in excel file to be made and then to
be opened in Jupyter notebook from ANACONDA Output of accessing data set is as follows:
Navigator.From there we can read the datasetin
ANACONDA Navigator. At Jupyter notebook
first we taking three variables for importing
purpose such as NumPy, matplotlib.pyplot and
pandas. These functions are used are as follows:
 NumPy is used for antialiasing a dynamic
array or large set.
 matplotlib.pyplot is used for making graph
 pandas are mainly used like database where
we store the dataset from the excel file.
Through the data set we are forming graph
and predicting
Then importing the dataset on the pandas through the
help of “<any variable>. read_csv” and then showing
that field that are importing from the excel file.
Through the salary sets the plotting the points at the Figure 1 Sample Dataset
graph. After then it is being tried to input linear
regression function makes only straight line, if all First we linear function. “LinearRegression()”
points are linear then these function is used and from these function it makes straight line through
directly goes making graph and predicting. But if the the points.
points are not linear so we go for next step. If the Output:
points are not linear then we use polynomial function Linear Regression
for curving purpose that help to observing user that
how the growth are moving on a path. After this the
diagram on the graph is for demonstration. Then it is
predicted to a salary through the graph using both the
x-y-axis.
“<variable>.predict(poly_reg.fit_transform(v))”.
From here it gets a prediction value through the
polynomial/linear regression.

IV. RESULTS AND DISCUSSION Figure 2: Results of linear regression for the salary
dataset
From this application can be observed as the
Now, plotting the point where x-axis represents the
graphical representation and can also predict the any
position and y-axis presents the salary from
point from position and automatically calculates database Position_Salaries.
salary. Also survey the salary venue.
Output:
Surveys of salary are therefore differentiated on basis of
their data source into those that -

Electronic copy available at: https://ssrn.com/abstract=3990877

 Sampling V. CONCLUSIONS
Choosing the salary from x-y graph as it is to use to
represent a specific data set that takes some trial and
error. Often there may be more than one appropriate
type of graph to be used. It will depend on the way of
choosing the way to present the data, as well as its
own preferences. Now a days spreadsheet programs
like Excel are very flexible to create graphs of
different types; with a few number of clicks one can
see the represented data as a bar graph as well as by
Figure 3: Sampling the data from the salary dataset a line graph, or a circle graph. This prediction is
correct upto a certain percentage. More accuracy can
In Figure 3 we can see all the points are not using. be obtained by implementing k-nearest regression.
For that reason now we are implementing From there the best prediction can be chosen. It can
“PolynomialFeatures” for curveness. “degree = 6” be improved in following way:
here refer the smoothness of curve . • It can give more advance software for tallying
Same like above here, salary medium.
“lg2.predict(poly_reg.fit_transform(x))” function is • It will host the platform on online servers.
used for curving than straight line. • It can do as large database also and curve also
Output: bigger than above example.
Polynomial Regression
This predictor method can be used for predicting
population of a country as well as forecasting a daily
issue.

Figure 4: Results of polynomial regression for the salary dataset

Electronic copy available at: https://ssrn.com/abstract=3990877

 REFERENCES
1. Andreas Mullar, “Introduction to
Machine Learning using Python: A guide
for data Scientist,” in O’Reilly Publisher,
India.
2. S. Marsland, Machine learning: an
algorithmic perspective. CRC press,
2015.
3. A. L. Buczak and E. Guven, “A survey of
data mining and machine learning
methods for cyber security intrusion
detection,” IEEE Communications
Surveys & Tutorials, vol. 18, no. 2, pp.
1153–1176, Oct., 2015.
4. Tzanis, George, et al. "Modern
Applications of Machine Learning."
Proceedings of the 1st Annual SEERC
Doctoral Student Conference–DSC.
2006.
5. Horvitz, Eric. "Machine learning,
reasoning, and intelligence in daily life:
Directions and challenges." Proceedings
of. Vol. 360. 2006.
6. Mitchell, Tom Michael. The discipline of
machine learning. Carnegie Mellon
University, School of Computer Science,
Machine Learning Department, 2006.
7. Arum, R. (1998). The effects of resources
on vocational student educational
outcomes: Invested dollars or diverted
dreams? Sociology of Education, 71, 130-
151.
8. Lewis, C. D., 1982. Industrial and
Business Forecasting Methods, London,
Butterworths.

Author Biographical Statements

Electronic copy available at: https://ssrn.com/abstract=3990877