Behaviour analysis of Amazon customer using novel POS–NEG composition‐based Pythagorean fuzzy relation

• Published date: 01 February 2022
• Author: Ajith Sundaram
• Date: 01 February 2022
• DOI: http://doi.org/10.1002/pa.2297

ACADEMIC PAPER

Behaviour analysis of Amazon customer using novel POS–NEG

composition-based Pythagorean fuzzy relation

Ajith Sundaram

1,2

1

Rajagiri Business School, Cochin, India

2

Department of Management, Amrita Vishwa

Vidyapeetham, Kochi, India

Correspondence

Ajith Sundaram, Rajagiri Business School,

Cochin, India.

Email: ajithsundaram@gmail.com

[Correction Statement: Correction added on

16 March 2022 after first online publication:

The second affiliation has been added in this

version.]

Uncertainty and vagueness are the important factors in any decision-making process.

Different tools and approaches had been introduced to handle the uncertainty and

vagueness environment of behaviour analysis. One of the latest tools in dealing with

uncertainty and vagueness is Pythagorean fuzzy sets which are the generalized form

of intuitionistic fuzzy set. In this paper, we explore the concept of Pythagorean fuzzy

sets and introduce POS–NEG composition to determine the behaviour of customers

in Amazon E-commerce website. The role of this enhanced composition is to gather

the sentiment of customers about the Amazon products. Our proposed work handles

multi-attribute, reduces complexity of the analysis and it is compared and proved to

be accurate than fuzzy set and intuitionistic fuzzy set. The main feature of Pythago-

rean fuzzy sets is its relatively novel mathematical framework in the fuzzy family with

higher ability to cope imprecision imbedded in decision-making.

1|INTRODUCTION

Reviews from consumers are very important in E-commerce systems.

Many online shops have developed review system for users to post

their reviews. With the rapid development of online E-commerce

websites, more and more people are willing to share their feelings,

opinions and suggestions. Customer feedback is the information pro-

vided by clients about whether they are satisfied or dissatisfied with a

product or service and about general experience they had with a com-

pany. Their opinion is a resource for improving customer experience

and adjusting company's actions to their needs. This information can

be collected from different kinds of surveys (positive feedback and

negative feedback). There are two major reasons why customer

reviews are important. First, negative reviews provide a company with

an opportunity to learn and develop. Positive reviews can be used as

customer testimonials during lead acquisition. In fact, 92% of con-

sumers are more likely to purchase a product after reading a positive

online review.

The experimental dataset is collected from Amazon.com. It con-

tains 14.7 million reviews on 24 categories of products with 2.4 mil-

lion items from December 2018 to March 2019. The dataset includes

basic information about products, such as books, electronics, fashion,

mobile and dress. We gather the opinion of customers about the

products and segregate the opinion into five categories such as must

be, one dimension, attractive, indifferent and reverse. These catego-

ries are importance to the customer when they buy the products. The

services assigned to each category are security, customer preference,

offers, more options and reverse.

The contribution of this work is to analyse the customer behav-

iour using enhanced POS–NEG composition of Pythagorean fuzzy to

conclude the customer reviews of group of products. This work is

organized as follows. Section 2 describes the literature survey. In Sec-

tion 3, we describes Pythagorean fuzzy set (PFS) and how the pro-

posed POS–NEG composition-based PFS can be used for customer

behaviour analysis of Amazon product reviews. Section 4 addresses

the comparison of accuracy of decision-making by our proposed work

with two existing systems. In Section 5, the last section, we conclude

the paper and provide direction for future studies.

2|LITERATURE SURVEY

Zhang, Chen, and Lu (2017) proposed a sentiment analysis technique

using aspect sentiment collaborative filtering algorithm which is com-

bined with a fuzzy Kano model. Proposed algorithm focused on the

customer's desire and concentrated on the characteristics of the prod-

uct by making better sentiment analysis from the customer's purchase

records, and then examine the customer's importance for each

Received: 7 August 2019 Revised: 22 June 2020 Accepted: 15 July 2020

DOI: 10.1002/pa.2297

J Public Affairs. 2022;22:e2297. wileyonlinelibrary.com/journal/pa © 2020 John Wiley & Sons Ltd 1of6

https://doi.org/10.1002/pa.2297