Acceso usuarios
Jorge Alberto Achcar, Daniel Marcos de Godoy Godoy, Fábio Ferraz Junior Ferraz
This study considers a customer complaint dataset due to the technical services provided by a telecommunications company collected for 134 consecutive weeks from the first week of January 2018 up to the year 2019. The total count of weekly complaints is the sum of different causes, which characterizes compositional data. The data was analyzed assuming a Poisson regression model for the weekly total complaint count data in presence of a random factor and compositional models both under a Bayesian approach using existing MCMC (Monte Carlo Markov Chain) to get the posterior summaries of interest. The obtained results are of great importance to improve the service quality of the company.
