Publication: Modelling of Churn in Telecom Industry Using Machine Learning
Date
2020-02
Authors
Amir Rushaidi bin Mohd Esa
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Churn analysis in telecom industry has long been a one of the major researches in
telecom companies in order for them to analyze their customer’s potential whether to
stop their subscription or keep the subscription. A huge and vast amount of information
are needed to analyze them to get the prediction as accurate as possible. Hence, machine
learning model are one of the effective methods required to help to further the
analyzation and achieved the desired results. This paperwork proposes a churn analysis
or prediction model that uses several methods and steps to make the prediction based
on a set of parameters from customers to know their potential to churn or not. The
analyzation is mainly conducted on WEKA Explorer Toolkit, a software developed by
the University of Waikato. From the raw dataset obtained, the file format is converted
into different file format, which then will be undergoing attribute selection process
using a method called Correlation Based Feature Selection. Finally, after the attribute
selection, customer classification process is executed to finally classify the class of the
customer which is narrowed down to two classes; churn or not churn. This paper also
introduces several other machine learning models that are used to compare between
other models on which will yield the best results based on several factors.
Description
FYP 2 SEM 2 2019/2020
Keywords
Modelling Churn Telecom