Análisis de la capacidad de las frecuencias 2.4ghz y 5ghz utilizando inteligencia artificial para la corrección de fallos en el servicio de internet doméstico proporcionado por la empresa Fiber Store

Abstract

This project establishes the relationship between Wi-Fi networks and artificial intelligence, after analyzing the information collected from the review of previous research works and the instruments of data acquisition applied to customers and technicians, some problems in the home networks were detected such as intermittent Wi-Fi signal, low connection speed, high packages loss rate, saturated bandwidth, high jitter and latency for the 2.4 GHz and 5 GHz frequency. The prediction model “Decision Tree Algorithm for Regression and Classification” is used because it was the model that best fitted the software requirements. This model allows measurements, network analysis, optimal values predictions, and failure detection carried out through the comparison of data obtained in the measurement and the values predicted by the algorithm, contemplating the conditions and limits established based on the frequency used and the speed plan contracted by the customers together. The developed tool provides an option to gather reports once the scan is completed and it predicts the correct location of the router in the customer's home. A technical guide is included, which contains methods to mitigate the failures detected on Wi-Fi networks. Among the proposed corrections there are change of channel, device management and load balancing, which allows the network to be supervised in a technical way, reducing the time used by each technical support provided to customers.