Home // International Journal On Advances in Telecommunications, volume 15, numbers 3 and 4, 2022 // View article

Linking Radio Access Network QoE and QoS with Ensemble Multiple Regression

Authors:
Adrien Schaffner
Louise Travé-Massuyès
Simon Pachy
Bertrand Le Marec

Keywords: Ensemble learning; Regression models; Data analysis; Knowledge extraction; Radio access networks; QoS/QoE relationship; Quality via QoE and customer reports

Abstract:
The evaluation of user satisfaction is an essential performance indicator for network operators. It can be impacted by several causes, including causes linked to the network. However, linking the subjective comments of a customer with an objective behavior of the network is an issue. Experience shows that an indicator taken from customer complaints gives a good trend on the level of network quality perceived by customers, but it is difficult to transpose into concrete actions because it is often unrelated to the key performance indicators on which engineers base their action plans. The objective of this work is to learn a model that links the complaint rate, expressed by the Customer Satisfaction Rate indicator, with a set of key performance indicators so that performance engineers better understand customer expectations and act foremost on the indicators that give the most dissatisfaction. To this end, this paper takes advantage of ensemble learning applied to multiple regression, focusing the ensemble strategy on variable selection. The model hence makes it possible to link Quality of Experience and Quality of Service, which is demonstrated by nice interpretable results obtained from applying the method to data from a French telecom case study.

Pages: 70 to 80

Copyright: Copyright (c) to authors, 2022. Used with permission.

Publication date: December 31, 2022

Published in: journal

ISSN: 1942-2601