Factores que caracterizan el tráfico gris en compañías de telefonía móvil

Abstract

Honduras currently suffers a deficiency in detection of mobile phone lines used for bypass fraud in GSM networks, due to the variable nature of the main factors considered, and the inflexibility of traditional detection methods. This study aims to find determinant factors of telephone lines used for bypass fraud, in order to use this knowledge to detect these cases through the use of a neural network with supervised learning. The results of this study are presented separately for the qualitative and quantitative approach, for a better understanding of the most significant factors discovered. With the characterization of user profiles for both the bypass telephone line and the legal telephone line, the study proceeds to build a neural network model with 99% detection efficiency, and for applicability matters, the basic considerations that need to be taken into account are described for the subsequent construction of a bypass detection model.