LTE downlink throughput modeling using neural networks

With the advancements made in the field of telecommunication, the quality of service is increasing day by day whilst the user load is simultaneously increasing on the service providers. In order to keep up with the tougher standards, a fairly large amount of money is spend on resource allocation, majority of which often ends up unused and as a result end up being wasted. In this paper we used LTE data obtained on hourly basis for a period of 60 days from a Telecommunication company and carried out its quantitative analysis using deep neural nets. The results have shown stark difference in utilization of resources across rural and urban areas. Also we were able to obtain a handful of key features, which play major role in determining the quality of data transmission, and emphasis on theses could ensure better quality at lesser cost.