Model to Improve the Forecast of the Content Caching based Time-Series Analysis at the Small Base Station

In the new cellular systems (5G), the approach of caching content in the small Base Stations (sBS) is considered to be a suitable approach to improve the efficiency and to reduce the user perceived latency content delivery. Proactively serving estimated users demands, via caching at sBS is crucial due to storage limitations. But, it requires knowledge about the content popularity distribution, which is often not available in advance. Moreover, human behavior is predictable, and contents popularity are subject to fluctuations since mobile users with different interests connect to the caching entity over time and in different places. In this paper, we focus on the prediction of popularity evolution of video contents/files, based on the observation of past solicitations. We propose the FORECASTING schemes to manage this problem based on the time series model Seasonal AutoRegressive Integrated Moving Average (SARIMA) to interpret the temporal influence. The scheme is based on two algorithms in static and dynamic cases to manage future cache decisions. Several numerical results are given with comments that confirm the proposed idea.