A study on the prediction of service reliability of wireless telecommunication system via distribution regression.

The reliability of wireless telecommunication service has become a major concern for the operation and maintenance (O&M) departments of the major telecommunication service providers. Consequently, reliability prediction has assumed a pivotal role for O&M department to furnish dependable services while simultaneously curtailing operational costs. However, predicting the service reliability is a challenging issue mainly attribute to the mobility characteristics of the end-users in the network. In tandem with the formulation of service reliability metrics, this study introduces a linear distribution regression model under Gaussian distribution and a kernel-based distribution regression model tailored for bimodal scenarios. The empirical validation of these proposed metrics and methodologies is substantiated through case studies employing both simulated data and real-world datasets derived from a wireless telecommunication system within an urban district. The outcomes of these case studies demonstrate the efficacy and applicability of the proposed metrics and methods. • Service reliability prediction for telecommunication network is considered. • A linear distribution regression is applied under Gaussian distribution. • A kernel-based distribution regression is proposed for bimodal distribution. • An adaptive algorithm is introduced to select the appropriate method. [ABSTRACT FROM AUTHOR]