Abstract: In the recent days, online social networks (OSNs) are widely known as one of the largest platforms for the source of information and content that can be shared among people. As a socially oriented learning-based model has been proposed with a view to provide quality of experience (QoE) for social media services. This is for media content prefetching in order to improve the access delay reduction and to enhance OSN user’s satisfaction. With a wide spread of data-driven analysis during a period of fourteen months, over the Twitter traces on real-life from over 2,800 users, it is revealed that the social relationship has a wide impact on user’s social media behaviour. In order to represent this scenario, a social relationship of clusters for a large group of friends has been conducted, and further to expand a cluster-based machine learning model for socially-oriented learning based prefetch prediction. And then to predict users influence on the social media app, a usage-adaptive prefetch scheduling mechanism has been used by considering different users which might possess heterogeneous user’s app usage way. A framework has been proposed that can be evaluated using a trace-driven analysis on the social media. The data can be dumped onto a cloud for further analysis by using different machine learning approaches.
Keywords: Online Social Networks, Prefetching, Quality of Experience, Spice, Twidere, Usage Adaptive prefetching scheduling.
|
DOI:
10.17148/IJARCCE.2021.10616
