A Hybrid Approach to Social Network User Feed Generation

Abstract

Existing social networks handle generation of user activity feeds by utilizing different data distribution models. Different models optimize different aspects of feed generation such as user specificity, processing efficiency, resource utilization and latency. This paper propo*ses a hybrid model to handle this problem elegantly. This model takes into account the frequency of query requests between individual users and classifies them into either a PUSH-Target user or PULLTarget user. The former is provided with prioritized data pushes and the latter with data pulls on user request basis.