DeepRANKPI: Time Series KPIs Prediction in a Live Cellular Network with RNN

Predictive analytics is employed by telecommunications operators to gain valuable insights on the network performance and make data-driven decisions in order to optimize the quality of service for the users while saving the network resources. However, predicting the network key performance indicators (KPIs) is a challenging task in a large scale cellular network with complex spatio-temporal variations. Moreover, to optimize, expand and modify the strategy of mobile network, many parameters are changed or features are being implemented through mobile networks every day. One of the major challenges in network maintenance is to track the side-effects of mentioned changes, in order to avoid any anomalies, by calculating the amount of degradation or improvement. Sometimes these changes coincide with other network seasonality behavior, hence, adding complications to the network. Inspired by the promising performance of Recurrent Neural Network (RNN) time-series prediction, we developed a KPI prediction model using RNN algorithm in a Mobile network with about 280000 cells of various technologies and frequencies. Model training is performed on three years of historic mobile network data. In this work, on one hand we have proposed methods to predict network KPIs in future interval, and on the other hand performance analysis of different methods are investigated.