Abstract

Quality of service (QoS) has been a main issue in the Nigerian Telecommunications industry, and how to improve it has been a challenge. The robustness of the KPIs is affected by meteorological variables like temperature, humidity, and rainfall. Intelligent monitoring systems that can anticipate the state of the KPIs and enable the policymakers in the telecommunication to take appropriate action before disruptions occur. These are necessary for mobile operators to lessen the effect of the meteorological parameters on KPIs and to also improve mobile services and user experience. Predictive models such as bagging, boosting (LSBoost), and Neural Networks (NN) were used to develop models for the key parameter indicators (KPIs) using meteorological parameters and evaluated using the mean absolute error (MAE) as the performance metric to be evaluated. The evaluation for Ado-Ekiti, Nigeria was evaluated using the MAE. After evaluation, it was concluded that the bagging model had the best performance for Ado-Ekiti mobile network designs out of the three models.