Publication detail

Understanding customer churn prediction research with structural topic models

FRIDRICH, M.

Original Title

Understanding customer churn prediction research with structural topic models

Type

journal article in Web of Science

Language

English

Original Abstract

Customer churn prediction is showing a growth in attention from both researchers and practitioners, creating a vast body of scientific works while being recognized as an indispensable tool of corporate retention activities. Thus, we aim to demonstrate the potential of structural topic models to navigate through the research articles and to identify essential themes and trends within the field of customer defection prediction. We apply a modified modeling procedure to journal articles focused on customer churn. As a result, the structural model of 38 topics is formed and examined considering topic prevalence, its changes over time, and the scientific impact (citations). We see prevailing themes tackling broad perspectives such as modeling, evaluation, and performance metrics. Furthermore, we recognize a slow decline in business & marketing aspects of churn prediction coupled with rising of more nuanced topics. At last, we discuss possible future steps in topic modeling within the domain.

Keywords

Customer Churn Prediction, Natural Language Processing, Topic Modeling

Authors

FRIDRICH, M.

Released

14. 12. 2020

Publisher

Academy of Economic Studies in Bucharest

Location

Bucharest, Romania

ISBN

1842-3264

Periodical

Economic Computation and Economic Cybernetics Studies and Research

Year of study

54

Number

4

State

Romania

Pages from

301

Pages to

317

Pages count

16

URL

BibTex

@article{BUT167274,
  author="Martin {Fridrich}",
  title="Understanding customer churn prediction research with structural topic models",
  journal="Economic Computation and Economic Cybernetics Studies and Research",
  year="2020",
  volume="54",
  number="4",
  pages="301--317",
  doi="10.24818/18423264/54.4.20.19",
  issn="1842-3264",
  url="http://ecocyb.ase.ro/nr2020_4/19.+Martin+FRIDRICH+(T).pdf"
}