Sequential dynamic event recommendation in event-based social networks: An upper confidence bound approach

In recent years, there have been some platforms that have focused on recommending commodities or events to users using event-based social networks (EBSNs). Some studies have attempted to find the optimal recommendation sequence of these items, assuming that the sequence stops once the user accepts one recommendation or the item list runs out. However, in reality, social media platforms will not stop recommending different commodities or social events to users until the user becomes bored and abandons the platform. Since it is 5 to 25 times more difficult to attract a new user than to retain an old one,' it would be helpful if the platform could determine when to stop making recommendations. In this work, we investigate the problem of sequential dynamic event recommendation with feedback (SDERF), where the platform continues recommending events even when the user has accepted one that is satisfactory. We first model the SDERF problem and provide two variants, namely, an online learning model with without contextual information. Then, we apply an upper confidence bound (UCB) approach with an expected regret polynomial in the number of events and rounds. Finally, we evaluate the performance of our proposed algorithms using both real and synthetic datasets. (C) 2020 Elsevier Inc. All rights reserved.

Automated summary

Recent studies have emphasized the importance of considering user feedback in recommending commodities or events on social media platforms. It is crucial for platforms to determine when to cease recommendations to prevent user attrition. By modeling the problem and applying the upper confidence bound approach, the performance of recommendation algorithms can be enhanced.