Abstract
Due to the increasing competitiveness in telecom’s market, it has now become more necessary foroperators to start building personal relationship with customers for targeted retention strategies. Achieving this goalrequires the development of an effective churn prediction model that will solve the problem of churn misclassification,which is persistent in current churn prediction models. With several existing segment-oriented churn predictionmodels failing to harness the power of associative networking provided by telecoms users, churn prediction accuracyremains unguaranteed while targeted decision support is not enhanced. Here, the research introduced the Customer’sInfluence Degree (I) to the existing Recency, Frequency, and Monetary (RFM) values as an additional predictivefactor, towards determining the churn class of a customer. The essence is to utilise the socio-transactional affinitiesof customers’ direct dependent to targeted communication nodes through customers RFM analysis to determinethe dominance of a customer in the community. The newly introduced predictive factor helped to minimise churnmisclassification rate through appropriate reclassification of customers who were wrongly classified as churner ornon-churner when using the existing RFM churn scores only.
Recommended Citation
Ibitoye, Ayodeji O.J; ONIME, Clement; Zaki, Nashwan Dheyaa; and Onifade, Olufade F.W
(2022)
"Socio-Transactional Impact of Recency, Frequency, and MonetaryFeatures oN Customers’ Behaviour in Telecoms’ Churn Prediction,"
Iraqi Journal for Computer Science and Mathematics: Vol. 3:
Iss.
2, Article 11.
DOI: https://doi.org/10.52866/ijcsm.2022.02.01.011
Available at:
https://ijcsm.researchcommons.org/ijcsm/vol3/iss2/11
