By Dr. Claudio Righetti, Emilia Gibellini, Florencia De Arca, Carlos Germán Carreño Romano, Mariela Fiorenzo, Gabriel Carro & Fernando Rodrigo Ochoa, Cablevisión S.A.
The purpose of this paper is to introduce STEM-ML, an extension of our network-dimensioning tool, which allows us to define the strategy to face the increasing demand, of both our Internet broadband and “Flow”, our Internet Protocol Television (IPTV) services. This tool makes use of machine learning techniques to characterize the optical nodes that integrate our network. Based on such characterization,we can define the technologic and commercial strategy for the access network so that Cablevisión (CVA) is able to afford the short and long-term demand.
Until the development of STEM-ML, characterization was made at hub level, and it was based on the average bandwidth per subscriber parameters. With STEM-ML, the analysis is made at optical node level, and monthly consumption, households passed (HHP), and protocol types, among other variables.
Moreover, data from “Flow” our IPTV platform is added. The increasing data volume generates the need for introducing machine learning and multivariate analysis techniques.