A PROJECT REPORT
ON
CUSTOMER SEGMENTATION USING MACHINE LEARNING
For the partial fulfillment for the award of the degree of
BACHELOR OF TECHNOLOGY
In
COMPUTER SCIENCE AND ENGINEERING
(AIML)
Submitted By
Ghanendra Singh (2101921539001)
Saif Riaz Khan (2101921539002)
Sameer Shekhar (2101921539003)
Tanmoy Chattopadhyay(2101921539004)
Under the Supervision of
Ms. ANJU CHANDNA
G.L. BAJAJ INSTITUTE OF TECHNOLOGY &
MANAGEMENT, GREATER NOIDA
Affiliated to
DR. APJ ABDUL KALAM TECHNICAL UNIVERSITY,
LUCKNOW
2022-23
1
TABLE OF CONTENT
Table of Content.............................................................................................................. 2
Declaration...................................................................................................................... 3
Certificate….................................................................................................................... 4
Acknowledgement .......................................................................................................... 5
Abstract…….................................................................................................................... 6
List of Figures …………………………………………………………………………. 7
Chapter 1. Introduction ........................................................................................... 8
1.1 General
1.2 Purpose
Chapter 2. Literature Survey .................................................................................. 10
Chapter 3. Methodology .......................................................................................... 12
Chapter 4. Software Requirements ......................................................................... 15
Chapter 5. Implementation...................................................................................... 16
Chapter 6. Result ....................................................................................................... 26
Chapter 7. Conclusion, Limitation & Future Scope ............................................... 28
Chapter 8. References ................................................................................................ 29
2
Declaration
We hereby declare that the project work presented in this report entitled “Customer
Segmentation using Machine Learning”, in partial fulfillment of the requirement for the
award of the degree of Bachelor of Technology in Computer Science & Engineering (AIML),
submitted to A.P.J. Abdul Kalam Technical University, Lucknow, is based on my own work
carried out at Department of Computer Science & Engineering, G.L. Bajaj Institute of
Technology & Management, Greater Noida. The work contained in the report is original and
project work reported in this report has not been submitted by me/us for award of any other
degree or diploma.
Signature: Signature:
Name: Ghanendra Singh Name: Saif Riaz Khan
Roll No: 2101921539001 Roll No: 2101921539002
Signature: Signature:
Name: Sameer Shekhar Name: Tanmoy Chattopadhyay
Roll No: 2101921539003 Roll No: 2101921539004
Date:
Place: Greater Noida
3
Certificate
This is to certify that the Project report entitled “Customer Segmentation using Machine
Learning” done by Ghanendra Singh (2101921539001), Saif Riaz Khan (2101921539002),
Sameer Shekhar (2101921539003) and Tanmoy Chattopadhyay (2101921539004) is an
original work carried out by them in Department of Computer Science & Engineering(AIML),
G.L Bajaj Institute of Technology & Management, Greater Noida under my guidance. The
matter embodied in this project work has not been submitted earlier for the award of any degree
or diploma to the best of my knowledge and belief.
Date:
Ms. Anju Chandna Dr. Sansar Singh Chauhan
Signature of the Supervisor Head of the Department
4
Acknowledgement
The merciful guidance bestowed to us by the almighty made us stick out this project to a
successful end. We humbly pray with sincere heart for his guidance to continue forever.
We pay thanks to our project guide Ms. Anju Chandna who has given guidance and light to
us during this project. Her versatile knowledge has cased us in the critical times during the span
of this project.
We pay special thanks to our Head of Department Dr. Sansar Singh Chauhan who has been
always present as a support and help us in all possible way during this project.
We also take this opportunity to express our gratitude to all those people who have been directly
and indirectly with us during the completion of the project.
We want to thanks our friends who have always encouraged us during this project.
At the last but not least thanks to all the faculty of CSE department who provided valuable
suggestions during the period of project.
5
Abstract
The emergence of many competitors and entrepreneurs has caused tons of tension between
competing businesses to find new buyers and keep old ones. As a result of the preceding, the
need for exceptional customer service becomes appropriate, regardless of the size of the
business. In addition, the ability of any business to understand each of its customers' needs will
receive greater support in providing targeted customer services and developing customized
customer service plans. This understanding is possible through structured customer service.
Each segment contains customers who share similar market features. Big data ideas and
machine learning have fostered more acceptance of the automated customer segmentation
approach in favor of traditional market analytics that often do not work especially when the
customer base is too large. In this paper, the k-Means clustering algorithm is used for this
purpose. The sklearn library was developed for the k-Means algorithm (found in the appendix)
and the program is trained using a two factor dataset of 100 patterns obtained from the retail
business. Features of the average number of customer purchases and the average number of
monthly customer visits.
6
LIST OF FIGURES
Page No.
Figure 3.1 15
Figure 5.1 18
Figure 5.2 21
Figure 5.3 22
Figure 5.4 23
Figure 5.5 24
Figure 6.1 27
Figure 6.2 28
7
Chapter 1
Introduction
1.1 General:
Over the years, the increasing competition between businesses and the availability of large-
scale historical data has resulted in the extensive use of data mining techniques to discover
important and strategic information that is hidden in the information of organizations. Data
mining is the process of extracting logical information from a dataset and presenting it in a
human-accessible way for decision support. Data mining techniques distinguish areas such
as statistics, artificial intelligence, machine learning and data systems. Data mining
applications include but are not limited to bioinformatics, weather forecasting, fraud
detection, financial analysis and customer segmentation. The key to this paper is to identify
customer segments in the commercial business using a data mining method. Customer
division is the division of the customer base of the business into groups called customer
segments such that each customer segment consists of customers who share similar market
characteristics. These distinctions are based on factors that can directly or indirectly
influence the market or business such as product preferences or expectations, locations,
behavior and so on. The importance of customer segmentation includes, inter alia, the ability
of a business to customize market plans that will be appropriate for each segment of its
customers; support for business decisions based on a risky environment such as debt
relations with their customers; Identification of products related to individual components
and how to manage demand and supply power; reveals the interdependence and interaction
between consumers, between products, or between customers and products that the business
may not be aware of; the ability to predict customer decline, and which customers are most
likely to have problems and raise other market research questions and provide clues to
finding solutions. Integrated proved effective for detecting subtle but subtle patterns or
relationships buried in a database of unencrypted data. This mode of learning is classified
under supervised learning. Integration algorithms include the k-Means algorithm, k-nearest
8
algorithm, Sorting Map (SOM) and more. These algorithms, without prior knowledge of the
data, are able to identify clusters in them by repeated comparisons of input patterns until
stable qualifications in the training examples are obtained depending on the subject matter
or the process. Each set contains data points that have very close similarities but vary greatly
from the data points of other clusters. Integration has great applications in pattern
recognition, image analysis, and bioinformatics and so on. In this paper, the k-Means
clustering algorithm was applied to the customer segment. The sklearn library (Appendix)
of the k-Means algorithm was developed, and the training was started using a standard
Silhouette -score with two feature sets of 100 training patterns found in the retail business.
After numerous indications, four stable intervals or customer segments were identified. Two
factors are considered in the combination of the number of goods purchased by the customer
per month and the average number of customer visits per month. From the dataset, four
customers or categories are grouped and labeled as follows: cluster_metrics_1,
cluster_metrics_2, cluster_metrics_3, cluster_metrics_4.
1.2 Purpose:
In this series of Data Science Project, we will make one of the most important applications
of machine learning - Customer Classification. For this project, we will use client
components in python. Whenever you need to find your best customer, customer division is
the best option. For this machine learning project, this project will provide you with a
background for customer segmentation. After that we will evaluate the data from which we
will build the classification model. Also, in this data science project, we will see a descriptive
analysis of our data and use several types of K-means algorithm. So, follow the complete
customer science project in the segment using an algorithm learning machine and run in
python. Customer Classification is one of the most important applications of unreadable
learning. Using merger strategies, companies can identify customer segments that allow
them to identify user bases. For this machine learning project, we will use K-methods
integration which is an important algorithm for integrating an unlisted dataset.
9
Chapter 2
LITERATURE SURVEY
A. Customer Classification:
Over the years, the commercial world has become more competitive, as organizations such
as these have to meet the needs and wants of their customers, attract new customers, and
thus improve their businesses. The task of identifying and meeting the needs and
requirements of each customer in the business is a very difficult task. This is because
customers may vary according to their needs, wants, demographics, shapes, taste and taste,
features and so on. As it is, it is a bad practice to treat all customers equally in business. This
challenge has led to the adoption of the concept of customer segmentation or market
segmentation, where consumers are divided into subgroups or segments where members of
each subcategory exhibit similar market behaviors or features. Accordingly, customer
segmentation is the process of dividing the market into indigenous groups.
B. Big Data:
Recently, Big Data research has gained momentum. defines big data as - a term that
describes a large number of formal and informal data, which cannot be analyzed using
traditional methods and algorithms. Companies include billions of data about their
customers, suppliers, and operations, and millions of internally connected sensors are sent
to the real world on devices such as mobile phones and cars, sensing, creating, and
communicating data. the ability to improve forecasting, save money, increase efficiency and
improve decision making in various fields such as traffic control, weather forecasting,
disaster prevention, finance, fraud control, business transactions, national security,
education, and healthcare. Big data is seen mainly in the three Vs namely: volume,
variability and speed. There are other 2Vs available - authenticity and value, thus making it
5V.
10
C. Data Collection:
Data collection is the process of collecting and measuring information against targeted
variations in an established system, enabling one to answer relevant questions and evaluate
results. Data collection is part of research in all fields of study including physical and social
sciences, humanities and business. The purpose of all data collection is to obtain quality
evidence that allows analysis to lead to the creation of convincing and misleading answers
to the questions submitted. We collected data from the UCI Machine Learning Repository.
D. Clustering Data:
Clustering is the process of grouping the information in the dataset based on some
similarities. There are a number of algorithms which can be chosen to be applied on a dataset
based on the situation provided. However, no universal clustering algorithm exists that's
why it becomes important to opt for appropriate clustering techniques. In this paper, we
have implemented three clustering algorithms using python sklearn library.
E. K-Mean:
K- means that an algorithm is one of the most popular classification algorithm. This
clustering algorithm depends on the centroid where each data point is placed in one of the
overlapping K clusters pre-programmed into the algorithm, The clusters are created that
correspond to the hidden pattern in the data that provides the information needed to help
decide the execution process. There are many ways to make k-means assembling; we will
use the elbow method.
11
Chapter 3
METHODOLOGY
The data used in this paper were collected from the UCI Machine Learning Repository. This
is a set of geographic data containing all transactions occurring between 1/1/2/10 and
9/12/2011 in an unregistered and unregistered UK broker. The company mainly sells unique
gifts all together. Many of the company's customers are shopkeepers. The database contains
8 attributes. These attributes include:
“Invoice No: Invoice number. By default, a 6-digit aggregate number is assigned separately
for each transaction. If this code starts with the letter 'c', it indicates the cancellation.”
Stock Code: Product (item). Name, a 5-digit number assigned only to each unique product.
”
“Definition: Product name (item). By name. ” “Price: The value of each product (item)
made. Number. "
“Invoice Date: Invitation Date and Time. In terms of numbers, the date and time of each
transaction. ”
“UnitPrice: Price is a unit. Prices, product price per unit of measurement."
“Customer: Customer number. Name, 5-digit number assigned to each customer. ”
“Country: Country name. Name, the name of the country where each customer lives. ”
In this paper several steps were taken to obtain an accurate result. It involves the addition
of a feature alongside the first step of the centroids, the allocation step and the update step,
which are the most common steps k-means algorithms.
A. Collect data
12
This is a data preparation phase. The feature usually helps to refine all data items at a
standard rate to improve the performance of the clustering algorithm. Each data point
changes from grade 2 to +2. Integration techniques that include Min-max, decimal and z-
points The standard z-signing strategy are used to make things unequal before applying the
k-Means algorithm to a dataset.
B. Customer Classification Methods
There are many ways to perform segmentation, which vary in severity, data requirements,
and purpose. The following are some of the most commonly used methods, but this is not
an incomplete list. There are papers that discuss artificial neural networks, particle fixation,
and complex types of ensemble, but are not included due to limited exposure. In future
articles, I may go into some of these alternatives, but for now, these more common methods
should be sufficient.
Each subsequent section of this article will include a basic description of the method, as
well as a code example for the method used. If you don't have the expertise, well, just skip
the code and you'll still have to get a good handle on each of the 4 sub-sections we include
in this article.
C. Group Analysis
Group analysis is a unifying, or unifying, approach for consumers based on their similarities.
There are 2 main types of group analysis categorized into market policy: Hierarchical group
analysis, and classification (Miller, 2015). In the meantime, we will discuss how to classify
clusters called k-methods.
D. K-means encounter
The k-means clustering algorithm is an algorithm that is frequently used to draw insights
into the formats and differences within a database. In marketing, it is often used to build
customer segments and to understand the behavior of these unique segments. Let's get into
building assembly models in the python environment.
E. Centroids Initiation
Selected cents or initials were selected. Figure 1 introduces the start of graduation centers.
Four selected centers shown in different shapes were selected using the Forgy method. In
13
Forgy's method of using k (in this case k = 4) data points are randomly selected as cluster
centroids.
14
Chapter 4
SOFTWARE REQUIREMENT
Hardware necessities
Hardware choice is essential to the standard and potency of any software package. In
Hardware choice, size and power necessities are necessary.
Customer isolation will be with success run on the system with AN i3 processor with a
minimum of four GB RAM and disc drive with 500GB and fifteen.6 inches to observe
system performance. (Printer is needed for text output).
• Pentium processor ------- 2 GHz or on top of
• RAM capability ------- 4 GB
• Hard Disk ------ 500 GB
Software necessities
One of the foremost troublesome tasks is, software package choice, as long because the
would like for the program is thought to search out out if a specific software package
package fits the wants. once the primary choice of alternatives safety is needed to urge the
need for a few software package compared to the opposite candidates. This section initial
summarizes the application's question so proposes an in depth comparison.
• Operating System : Windows seven or ten
• Software: Google Colab
• Databases: Excel sheets
• Python Libraries
15
Chapter 5
IMPLEMENTATION
Import packages and data
To start, we import the packages needed to do our analysis and then import the xlsx (excel
spreadsheet) data file. If you want to follow along with the same data, you'll need to
download it from UCI. For this example, I put the xlsx file in the folder (directory) where I
present Google Colab.
As you can see, we have 8 columns of data for each row and each row represents an item
purchased. This isn’t that helpful yet, so let’s clean and organize this data in a way that
allows us to formulate more actionable insights.
Data cleaning
Below, we will remove data that is not helpful, missing, or potentially cause issues in the
long run.
Now let's convert the data so that each record represents one customer purchase
history.
16
We now have a DataFrame with complete sales, order counts, and average order price
per customer. But right now we're not home.
Normalize the data
Clustering algorithms like K-means are sensitive to the scales of the data used, so
we’ll want to normalize the data.
Below is a screenshot from part of a Stack Exchange answer discussing why
standardization or normalization is necessary for data used in K-means clustering. The
screenshot is linked to the Stack Exchange question, so you can click on it and read the
entirety of the discussion if you’d like more information.
17
Our data is scaled between -2 and 2. Now let’s get to clustering.
Select the optimal number of clusters
Alright, we’re ready to run cluster analysis. But first, we need to figure out how many
clusters we want to use. There are several approaches to selecting the number of clusters
to use, but I’m going to cover two in this article: (1) silhouette coefficient, and (2) the
elbow method.
Silhouette (Clustering)
Silhouette means how to interpret and verify consistency within data structures. This
method provides a picture showing how well each item is organized. [1]
The value of a silhouette is a measure of how something is similar in its collections
(combinations) compared to other clusters (divisions). The silhouette goes from –1 to
+1, where a higher value indicates that an item is properly matched to its collection and
compared to neighboring clusters. If multiple objects have a high value, then the
integration configuration is appropriate. If most points have a value or a negative value,
then the coordinate system may have too many or too few clusters.
The silhouette can be calculated with any distance metric, such as Euclidean distance or
Manhattan distance.
18
Now that we know a whole lot more of the silhouette, let's go in and use the code to find
the right number of clusters.
Cluster 4 had the most complete silhouette fit, indicating that 4 could be the best number
of clusters. But we'll look at that twice with the elbow way.
Elbow Criterion Method(with the Sum of Squared Errors (SSE)):
The idea behind the elbow method is to run the k-mean correlation in the given data for
k values (num_clusters, e.g. k = 1 to 10), and for each k value, to calculate the sum of
squared errors (SSE).
After that, adjust the SSE line for each k value. If the line graph looks like an arm - a red
circle below the line of the line (as an angle), the "elbow" on the arm is the correct price
(collection value). Here, we want to reduce the SSE. SSE usually drops to 0 as we go up
k (and SSE is 0 where k equals the number of data points, because where each data point
is its own set, and there is no error between it and its trunk).
Therefore the purpose is to select a small value of k that still has a low SSE, and the cone
usually represents where it starts to have a negative return with increasing k.
19
Well, with the correct understanding of the elbow mechanism in hand, let's use the elbow
method to see if it agrees with our previous results suggesting 4 sets.
Based on the graph above, it looks like K = 4, or 4 clusters is the correct number of clusters
in this analysis. Now let's translate the customer segments provided by these components.
Interpreting Customer Segments
20
21
Now let's combine the metrics of the integration and see what we can gather from the
standard data for each cluster.
In the following section, we will visualize the clustering by adding different columns to
the x and y axis. Let's see what we say.
Green customers have the lowest price and lowest order count, which means they are the
lowest bidder. On the other hand, orange customers have the highest total SALE and
highest order count, indicating that they are the highest priced customer.
22
In this structure, we consider the average order value vs the order value. Once again, green
buyers are the lowest price and the customers in the orange are the highest prices.
You can look at it this way. You can target customers in red graphics and try to find ways
to increase their order count through email reminders or SMS notifications directed to
other identification features. Maybe you can email them a discount if they come back
within 30 days. Ideally, you can provide a delayed coupon (which will be used at some
point) at checkout.
23
Similarly, with customers who are in the blue segment, you may want to try other sales and
marketing strategies for the cart. Probably the fastest offer, based on market basket analysis
(see section on market basket analysis below).
In this building, it has a median value and order compared to the total retail price. This
structure also strengthens the previous 2 sites in identifying the orange group as the highest
value customers, the green as the lowest priced customers, and the blue and red as the high
potential customers.
From a growth perspective, I focus my attention on the blue and red collection. I try to better
understand each encounter and their intelligent behavior on site to identify which team to
focus on first and introduce a few test cycles.
The Best-selling item by segment
We know we have 4 categories and we know how much they spend on each purchase, their
total usage, and the number of their orders. The next thing we can do is to help us better
understand customer segments to find out which items are best sold in each segment.
24
Based on this information, we now know that the Jumbo Bag Red Retrospot is the best-
selling item by our most expensive team. With that information available, we can make
recommendations for other potential customers in this section.
25
Chapter 6
RESULT
26
27
Chapter 7
CONCLUSION
In this project, segments of customers are created using the k-means clustering model
and analyzed the dataset, in various ways. Visualization of the data set has been done
for the better understanding about all the elements and its relation between the data. We
used a clustering approach called K-means clustering, in particular. K-means clustering
is one of the most popular clustering methods, and it's frequently the first thing
practitioners try when they're working on a clustering problem. K- means are used to
divide data points into discrete, non-overlapping groupings. One of the most common
uses of K-means clustering is client segmentation in order to gain a better understanding
of them, which can then be used to boost the company's income.
28
Chapter 8
REFERENCES
[1] Blanchard, Tommy. Bhatnagar, Pranshu. Behera, Trash. (2019). Marketing Analytics
Scientific Data: Achieve your marketing objectives with Python's data analytics
capabilities. S.l: Packt printing is limited.
[2] Griva, A., Bardaki, C., Pramatari, K., Papakiriakopoulos, D. (2018). Sales business
analysis: Customer categories use market basket data. Systems Expert Systems, 100, 1-
16.
[3] Hong, T., Kim, E. (2011). It separates consumers from online stores based on factors
that affect the customer's intention to purchase. Expert System Applications, 39 (2),
2127-2131.
[4] Hwang, Y. H. (2019). Hands-on Advertising Science Data: Develop your machine
learning marketing strategies… using python and r. S.l: Packt printing is limited.
[5] Puwanenthiren Premkanth, - Market Classification and Its Impact on Customer
Satisfaction and Special Reference to the Commercial Bank of Ceylon PLC.‖ Global
Journal of Management and Business Publisher Research: Global Magazenals Inc.
(USA). 2012. Print ISSN: 0975-5853. Volume 12 Issue 1.
[6] Puwanenthiren Premkanth, - Market Classification and Its Impact on Customer
Satisfaction and Special Reference to the Commercial Bank of Ceylon PLC.‖ Global
Journal of Management and Business Publisher Research: Global Magazenals Inc.
(USA). 2012. Print ISSN: 0975-5853. Volume 12 Issue 1.
[7] Sulekha Goyat. "The basis of market segmentation: a critical review of the literature.
European Journal of Business and Management www.iiste.org. 2011. ISSN 2222- 1905
(Paper) ISSN 2222-2839 (Online). Vol 3, No.9, 2011.
[8] By Jerry W Thomas. 2007. Accessed at: www.decisionanalyst.com on July 12, 2015.
[9] T.Nelson Gnanaraj, Dr.K.Ramesh Kumar N.Monica. AnuManufactured cluster
29
analysis using a new algorithm from structured and unstructured data. International
Journal of Advances in Computer Science and Technology. 2007. Volume 3, No.2.
[10]McKinsey Global Institute. Big data. The next frontier is creativity, competition and
productivity. 2011. Accessed at: www.mckinsey.com/mgi on July 14,2015.
[11] Jean Yan. - Big Data, Big Opportunities- Domains of Data.gov: Promote, lead,
contribute, and collaborate in the big data era. 2013. Retrieved from:
http://www.meritalk.com/pdfs/bdx/bdxwhitepaper090413.pdf July 14, 2015.
30
